From e59a393d089d08a4622de07f941dd3629fcaec6a Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 9 Jun 2016 12:02:03 -0700
Subject: [PATCH 01/14] fm10k: fix PCI device enable_cnt leak in .io_slot_reset
A previous patch removed the pci_disable_device() call in
.io_error_detected. This call corresponded to a pci_enable_device_mem()
call within .io_slot_reset handler. Change the call here to
a pci_reenable_device() so that it does not increment and leak the
enable_cnt reference count for the device. Without this change, VF
devices may fail during an unbind/bind, and we'll never zero the
reference counter for the pci_dev structure.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index 774a5654bf424..377a3fbb779fc 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -2275,7 +2275,7 @@ static pci_ers_result_t fm10k_io_slot_reset(struct pci_dev *pdev)
{
pci_ers_result_t result;
- if (pci_enable_device_mem(pdev)) {
+ if (pci_reenable_device(pdev)) {
dev_err(&pdev->dev,
"Cannot re-enable PCI device after reset.\n");
result = PCI_ERS_RESULT_DISCONNECT;
From 5b9e4432db038eefcafe2be468e3adaf2edbbe91 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 9 Jun 2016 14:56:05 -0700
Subject: [PATCH 02/14] fm10k: use software values when checking for Tx hangs
in hot path
A previous patch added support to check for hardware Tx pending in the
fm10k_down routine. This support was intended to ensure that we
accurately check what the hardware state is. However, checking for Tx
hangs in this manor during the hotpath results in a large performance
hit. Avoid this by making the hotpath check use the SW counters instead.
Fixes: a0f53cf49cb0 ("fm10k: use actual hardware registers when checking for pending Tx", 2016-06-08)
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k.h | 2 +-
drivers/net/ethernet/intel/fm10k/fm10k_main.c | 19 +++++++++++++++----
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 2 +-
3 files changed, 17 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k.h b/drivers/net/ethernet/intel/fm10k/fm10k.h
index c4cf08dcf5af9..75429f2bbc972 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k.h
@@ -458,7 +458,7 @@ __be16 fm10k_tx_encap_offload(struct sk_buff *skb);
netdev_tx_t fm10k_xmit_frame_ring(struct sk_buff *skb,
struct fm10k_ring *tx_ring);
void fm10k_tx_timeout_reset(struct fm10k_intfc *interface);
-u64 fm10k_get_tx_pending(struct fm10k_ring *ring);
+u64 fm10k_get_tx_pending(struct fm10k_ring *ring, bool in_sw);
bool fm10k_check_tx_hang(struct fm10k_ring *tx_ring);
void fm10k_alloc_rx_buffers(struct fm10k_ring *rx_ring, u16 cleaned_count);
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index e9767b6366a83..0d6e69d1586a6 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -1128,13 +1128,24 @@ static u64 fm10k_get_tx_completed(struct fm10k_ring *ring)
return ring->stats.packets;
}
-u64 fm10k_get_tx_pending(struct fm10k_ring *ring)
+/**
+ * fm10k_get_tx_pending - how many Tx descriptors not processed
+ * @ring: the ring structure
+ * @in_sw: is tx_pending being checked in SW or in HW?
+ */
+u64 fm10k_get_tx_pending(struct fm10k_ring *ring, bool in_sw)
{
struct fm10k_intfc *interface = ring->q_vector->interface;
struct fm10k_hw *hw = &interface->hw;
+ u32 head, tail;
- u32 head = fm10k_read_reg(hw, FM10K_TDH(ring->reg_idx));
- u32 tail = fm10k_read_reg(hw, FM10K_TDT(ring->reg_idx));
+ if (likely(in_sw)) {
+ head = ring->next_to_clean;
+ tail = ring->next_to_use;
+ } else {
+ head = fm10k_read_reg(hw, FM10K_TDH(ring->reg_idx));
+ tail = fm10k_read_reg(hw, FM10K_TDT(ring->reg_idx));
+ }
return ((head <= tail) ? tail : tail + ring->count) - head;
}
@@ -1143,7 +1154,7 @@ bool fm10k_check_tx_hang(struct fm10k_ring *tx_ring)
{
u32 tx_done = fm10k_get_tx_completed(tx_ring);
u32 tx_done_old = tx_ring->tx_stats.tx_done_old;
- u32 tx_pending = fm10k_get_tx_pending(tx_ring);
+ u32 tx_pending = fm10k_get_tx_pending(tx_ring, true);
clear_check_for_tx_hang(tx_ring);
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index 377a3fbb779fc..20c2a0cf83da2 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -1699,7 +1699,7 @@ void fm10k_down(struct fm10k_intfc *interface)
/* start checking at the last ring to have pending Tx */
for (; i < interface->num_tx_queues; i++)
- if (fm10k_get_tx_pending(interface->tx_ring[i]))
+ if (fm10k_get_tx_pending(interface->tx_ring[i], false))
break;
/* if all the queues are drained, we can break now */
From 4aa0bd54d4234d7bdc7cce331e7097c9393aca6e Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 9 Jun 2016 15:42:36 -0700
Subject: [PATCH 03/14] fm10k: use variadic form of alloc_workqueue
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index 0d6e69d1586a6..316f95b2b6004 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -56,7 +56,7 @@ static int __init fm10k_init_module(void)
pr_info("%s\n", fm10k_copyright);
/* create driver workqueue */
- fm10k_workqueue = alloc_workqueue("fm10k", WQ_MEM_RECLAIM, 0);
+ fm10k_workqueue = alloc_workqueue("%s", WQ_MEM_RECLAIM, 0, fm10k_driver_name);
fm10k_dbg_init();
From 88cdcfec9a46af47291207d86493e919093df3d1 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Fri, 17 Jun 2016 14:36:45 -0700
Subject: [PATCH 04/14] fm10k: remove fm10k_get_reta_size from namespace
The function is only used in fm10k_ethtool.c, so make it static.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k.h | 1 -
drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c | 2 +-
2 files changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k.h b/drivers/net/ethernet/intel/fm10k/fm10k.h
index 75429f2bbc972..e9850697be043 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k.h
@@ -496,7 +496,6 @@ int fm10k_close(struct net_device *netdev);
/* Ethtool */
void fm10k_set_ethtool_ops(struct net_device *dev);
-u32 fm10k_get_reta_size(struct net_device *netdev);
void fm10k_write_reta(struct fm10k_intfc *interface, const u32 *indir);
/* IOV */
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c b/drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c
index c04cbe9c9f7ca..adb7cb4311ba1 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c
@@ -966,7 +966,7 @@ static int fm10k_set_priv_flags(struct net_device *netdev, u32 priv_flags)
return 0;
}
-u32 fm10k_get_reta_size(struct net_device __always_unused *netdev)
+static u32 fm10k_get_reta_size(struct net_device __always_unused *netdev)
{
return FM10K_RETA_SIZE * FM10K_RETA_ENTRIES_PER_REG;
}
From ce4dad2ce231aa5258ddfc98f8a80d958643c014 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Fri, 17 Jun 2016 16:21:11 -0700
Subject: [PATCH 05/14] fm10k: prefer READ_ONCE instead of ACCESS_ONCE
While technically not needed, as all our uses of ACCESS_ONCE are scalar
types, we already use READ_ONCE in a few places, and for code
readability we can swap all the uses of the older ACCESS_ONCE into
READ_ONCE.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_common.h | 4 ++--
drivers/net/ethernet/intel/fm10k/fm10k_iov.c | 4 ++--
drivers/net/ethernet/intel/fm10k/fm10k_main.c | 2 +-
drivers/net/ethernet/intel/fm10k/fm10k_netdev.c | 6 +++---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 2 +-
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_common.h b/drivers/net/ethernet/intel/fm10k/fm10k_common.h
index 50f71e997448e..d51f9c7a47ff4 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_common.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_common.h
@@ -34,7 +34,7 @@ u32 fm10k_read_reg(struct fm10k_hw *hw, int reg);
/* write operations, indexed using DWORDS */
#define fm10k_write_reg(hw, reg, val) \
do { \
- u32 __iomem *hw_addr = ACCESS_ONCE((hw)->hw_addr); \
+ u32 __iomem *hw_addr = READ_ONCE((hw)->hw_addr); \
if (!FM10K_REMOVED(hw_addr)) \
writel((val), &hw_addr[(reg)]); \
} while (0)
@@ -42,7 +42,7 @@ do { \
/* Switch register write operations, index using DWORDS */
#define fm10k_write_sw_reg(hw, reg, val) \
do { \
- u32 __iomem *sw_addr = ACCESS_ONCE((hw)->sw_addr); \
+ u32 __iomem *sw_addr = READ_ONCE((hw)->sw_addr); \
if (!FM10K_REMOVED(sw_addr)) \
writel((val), &sw_addr[(reg)]); \
} while (0)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_iov.c b/drivers/net/ethernet/intel/fm10k/fm10k_iov.c
index 47f0743ec03b8..d9dec81f6b6d0 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_iov.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_iov.c
@@ -51,7 +51,7 @@ s32 fm10k_iov_event(struct fm10k_intfc *interface)
int i;
/* if there is no iov_data then there is no mailbox to process */
- if (!ACCESS_ONCE(interface->iov_data))
+ if (!READ_ONCE(interface->iov_data))
return 0;
rcu_read_lock();
@@ -99,7 +99,7 @@ s32 fm10k_iov_mbx(struct fm10k_intfc *interface)
int i;
/* if there is no iov_data then there is no mailbox to process */
- if (!ACCESS_ONCE(interface->iov_data))
+ if (!READ_ONCE(interface->iov_data))
return 0;
rcu_read_lock();
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index 316f95b2b6004..4b65364fb322e 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -1408,7 +1408,7 @@ static void fm10k_update_itr(struct fm10k_ring_container *ring_container)
* that the calculation will never get below a 1. The bit shift
* accounts for changes in the ITR due to PCIe link speed.
*/
- itr_round = ACCESS_ONCE(ring_container->itr_scale) + 8;
+ itr_round = READ_ONCE(ring_container->itr_scale) + 8;
avg_wire_size += BIT(itr_round) - 1;
avg_wire_size >>= itr_round;
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
index 20a5bbe3f5361..855833e85aacc 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
@@ -1098,7 +1098,7 @@ static struct rtnl_link_stats64 *fm10k_get_stats64(struct net_device *netdev,
rcu_read_lock();
for (i = 0; i < interface->num_rx_queues; i++) {
- ring = ACCESS_ONCE(interface->rx_ring[i]);
+ ring = READ_ONCE(interface->rx_ring[i]);
if (!ring)
continue;
@@ -1114,7 +1114,7 @@ static struct rtnl_link_stats64 *fm10k_get_stats64(struct net_device *netdev,
}
for (i = 0; i < interface->num_tx_queues; i++) {
- ring = ACCESS_ONCE(interface->tx_ring[i]);
+ ring = READ_ONCE(interface->tx_ring[i]);
if (!ring)
continue;
@@ -1299,7 +1299,7 @@ static void *fm10k_dfwd_add_station(struct net_device *dev,
static void fm10k_dfwd_del_station(struct net_device *dev, void *priv)
{
struct fm10k_intfc *interface = netdev_priv(dev);
- struct fm10k_l2_accel *l2_accel = ACCESS_ONCE(interface->l2_accel);
+ struct fm10k_l2_accel *l2_accel = READ_ONCE(interface->l2_accel);
struct fm10k_dglort_cfg dglort = { 0 };
struct fm10k_hw *hw = &interface->hw;
struct net_device *sdev = priv;
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index 20c2a0cf83da2..11b29552f6646 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -62,7 +62,7 @@ u16 fm10k_read_pci_cfg_word(struct fm10k_hw *hw, u32 reg)
u32 fm10k_read_reg(struct fm10k_hw *hw, int reg)
{
- u32 __iomem *hw_addr = ACCESS_ONCE(hw->hw_addr);
+ u32 __iomem *hw_addr = READ_ONCE(hw->hw_addr);
u32 value = 0;
if (FM10K_REMOVED(hw_addr))
From e5fbfb78641ff0c5139ae665289ed9f91524265e Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Mon, 20 Jun 2016 10:39:32 -0700
Subject: [PATCH 06/14] fm10k: NAPI polling routine must return actual work
done
When fm10k_poll fully cleans rings it returns 0. This is incorrect as it
messes up the budget accounting in the core NAPI code. Fix this by
returning actual work done, capped at budget - 1 since the core doesn't
expect a return of the full budget when the driver modifies the NAPI
status.
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Venkatesh Srinivas <venkateshs@google.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index 4b65364fb322e..32d44290b04f1 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -1484,7 +1484,7 @@ static int fm10k_poll(struct napi_struct *napi, int budget)
/* re-enable the q_vector */
fm10k_qv_enable(q_vector);
- return 0;
+ return min(work_done, budget - 1);
}
/**
From 76ef0fc5a7519690a6a87a248f5709463f584292 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 23 Jun 2016 13:31:00 -0700
Subject: [PATCH 07/14] fm10k: print error code when pci_enable_device_mem
fails during probe
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index 11b29552f6646..e5f37b7881968 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -1951,8 +1951,11 @@ static int fm10k_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
int err;
err = pci_enable_device_mem(pdev);
- if (err)
+ if (err) {
+ dev_err(&pdev->dev,
+ "PCI enable device failed: %d\n", err);
return err;
+ }
err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(48));
if (err)
From 18095937cb1c66b8ff944de02cf04d1497008352 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 23 Jun 2016 13:31:01 -0700
Subject: [PATCH 08/14] fm10k: don't continue probe if PCI device not in normal
IO state
In the event of an uncorrectable AER error occurring when the driver has
not loaded, the recovery routines are not done. This is done because
future loads of the driver may not be aware of the IO state and may not
be able to recover at all. In this case, when we next load the driver it
fails due to what appears to be a surprise remove event. Instead, add
a check to ensure that the device is in the normal IO state before
continuing to probe. This allows us to give a more descriptive message
of what is wrong.
Without this change, the driver will attempt to probe up to our first
call of .reset_hw() which will be unable to read registers and act as if
a surprise remove event occurred.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index e5f37b7881968..860fe04f4c72b 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -1950,6 +1950,12 @@ static int fm10k_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
struct fm10k_intfc *interface;
int err;
+ if (pdev->error_state != pci_channel_io_normal) {
+ dev_err(&pdev->dev,
+ "PCI device still in an error state. Unable to load...\n");
+ return -EIO;
+ }
+
err = pci_enable_device_mem(pdev);
if (err) {
dev_err(&pdev->dev,
From 5f45c83024c49f71e50cf8156b165f11d2ab35a9 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 23 Jun 2016 13:54:01 -0700
Subject: [PATCH 09/14] fm10k: don't try to stop queues if we've lost hw_addr
In the event of a surprise remove, we expect the driver to go down,
which includes calling .stop_hw(). However, this function will return an
error because the queues won't appear to cleanly disable. Prevent this
and avoid the unnecessary checks by just returning when
FM10K_REMOVED(hw->hw_addr) is true.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_common.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_common.c b/drivers/net/ethernet/intel/fm10k/fm10k_common.c
index d6baaea8bc7c1..dd95ac4f4c64a 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_common.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_common.c
@@ -207,6 +207,9 @@ s32 fm10k_disable_queues_generic(struct fm10k_hw *hw, u16 q_cnt)
/* clear tx_ready to prevent any false hits for reset */
hw->mac.tx_ready = false;
+ if (FM10K_REMOVED(hw->hw_addr))
+ return 0;
+
/* clear the enable bit for all rings */
for (i = 0; i < q_cnt; i++) {
reg = fm10k_read_reg(hw, FM10K_TXDCTL(i));
From f92e0e489225cae41290f8b42bf04128b2451789 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Fri, 26 Aug 2016 00:14:34 -0700
Subject: [PATCH 10/14] fm10k: rework vxlan_port offload before adding geneve
support
In preparation for adding Geneve Rx offload support, refactor the
current VXLAN offload flow to be a bit more generic so that it will be
easier to add the new Geneve code. The fm10k hardware supports one VXLAN
and one Geneve tunnel, so we will eventually treat the VXLAN and Geneve
tunnels identically. To this end, factor out the code that handles the
current list so that we can use the generic flow for both tunnels in the
next patch.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k.h | 6 +-
drivers/net/ethernet/intel/fm10k/fm10k_main.c | 4 +-
.../net/ethernet/intel/fm10k/fm10k_netdev.c | 161 ++++++++++--------
3 files changed, 93 insertions(+), 78 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k.h b/drivers/net/ethernet/intel/fm10k/fm10k.h
index e9850697be043..209268a147120 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k.h
@@ -240,9 +240,7 @@ struct fm10k_iov_data {
struct fm10k_vf_info vf_info[0];
};
-#define fm10k_vxlan_port_for_each(vp, intfc) \
- list_for_each_entry(vp, &(intfc)->vxlan_port, list)
-struct fm10k_vxlan_port {
+struct fm10k_udp_port {
struct list_head list;
sa_family_t sa_family;
__be16 port;
@@ -335,7 +333,7 @@ struct fm10k_intfc {
u32 reta[FM10K_RETA_SIZE];
u32 rssrk[FM10K_RSSRK_SIZE];
- /* VXLAN port tracking information */
+ /* UDP encapsulation port tracking information */
struct list_head vxlan_port;
#ifdef CONFIG_DEBUG_FS
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index 32d44290b04f1..0d39103124bd3 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -651,11 +651,11 @@ static int fm10k_clean_rx_irq(struct fm10k_q_vector *q_vector,
static struct ethhdr *fm10k_port_is_vxlan(struct sk_buff *skb)
{
struct fm10k_intfc *interface = netdev_priv(skb->dev);
- struct fm10k_vxlan_port *vxlan_port;
+ struct fm10k_udp_port *vxlan_port;
/* we can only offload a vxlan if we recognize it as such */
vxlan_port = list_first_entry_or_null(&interface->vxlan_port,
- struct fm10k_vxlan_port, list);
+ struct fm10k_udp_port, list);
if (!vxlan_port)
return NULL;
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
index 855833e85aacc..07c5a685f0e94 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
@@ -384,129 +384,147 @@ static void fm10k_request_glort_range(struct fm10k_intfc *interface)
}
/**
- * fm10k_del_vxlan_port_all
+ * fm10k_free_udp_port_info
* @interface: board private structure
*
* This function frees the entire vxlan_port list
**/
-static void fm10k_del_vxlan_port_all(struct fm10k_intfc *interface)
+static void fm10k_free_udp_port_info(struct fm10k_intfc *interface)
{
- struct fm10k_vxlan_port *vxlan_port;
-
- /* flush all entries from list */
- vxlan_port = list_first_entry_or_null(&interface->vxlan_port,
- struct fm10k_vxlan_port, list);
- while (vxlan_port) {
- list_del(&vxlan_port->list);
- kfree(vxlan_port);
- vxlan_port = list_first_entry_or_null(&interface->vxlan_port,
- struct fm10k_vxlan_port,
- list);
+ struct fm10k_udp_port *port;
+
+ /* flush all entries from vxlan list */
+ port = list_first_entry_or_null(&interface->vxlan_port,
+ struct fm10k_udp_port, list);
+ while (port) {
+ list_del(&port->list);
+ kfree(port);
+ port = list_first_entry_or_null(&interface->vxlan_port,
+ struct fm10k_udp_port,
+ list);
}
}
/**
- * fm10k_restore_vxlan_port
+ * fm10k_restore_udp_port_info
* @interface: board private structure
*
- * This function restores the value in the tunnel_cfg register after reset
+ * This function restores the value in the tunnel_cfg register(s) after reset
**/
-static void fm10k_restore_vxlan_port(struct fm10k_intfc *interface)
+static void fm10k_restore_udp_port_info(struct fm10k_intfc *interface)
{
struct fm10k_hw *hw = &interface->hw;
- struct fm10k_vxlan_port *vxlan_port;
+ struct fm10k_udp_port *port;
/* only the PF supports configuring tunnels */
if (hw->mac.type != fm10k_mac_pf)
return;
- vxlan_port = list_first_entry_or_null(&interface->vxlan_port,
- struct fm10k_vxlan_port, list);
+ port = list_first_entry_or_null(&interface->vxlan_port,
+ struct fm10k_udp_port, list);
/* restore tunnel configuration register */
fm10k_write_reg(hw, FM10K_TUNNEL_CFG,
- (vxlan_port ? ntohs(vxlan_port->port) : 0) |
+ (port ? ntohs(port->port) : 0) |
(ETH_P_TEB << FM10K_TUNNEL_CFG_NVGRE_SHIFT));
}
+static struct fm10k_udp_port *
+fm10k_remove_tunnel_port(struct list_head *ports,
+ struct udp_tunnel_info *ti)
+{
+ struct fm10k_udp_port *port;
+
+ list_for_each_entry(port, ports, list) {
+ if ((port->port == ti->port) &&
+ (port->sa_family == ti->sa_family)) {
+ list_del(&port->list);
+ return port;
+ }
+ }
+
+ return NULL;
+}
+
+static void fm10k_insert_tunnel_port(struct list_head *ports,
+ struct udp_tunnel_info *ti)
+{
+ struct fm10k_udp_port *port;
+
+ /* remove existing port entry from the list so that the newest items
+ * are always at the tail of the list.
+ */
+ port = fm10k_remove_tunnel_port(ports, ti);
+ if (!port) {
+ port = kmalloc(sizeof(*port), GFP_ATOMIC);
+ if (!port)
+ return;
+ port->port = ti->port;
+ port->sa_family = ti->sa_family;
+ }
+
+ list_add_tail(&port->list, ports);
+}
+
/**
- * fm10k_add_vxlan_port
+ * fm10k_udp_tunnel_add
* @netdev: network interface device structure
* @ti: Tunnel endpoint information
*
- * This function is called when a new VXLAN interface has added a new port
- * number to the range that is currently in use for VXLAN. The new port
- * number is always added to the tail so that the port number list should
- * match the order in which the ports were allocated. The head of the list
- * is always used as the VXLAN port number for offloads.
+ * This function is called when a new UDP tunnel port has been added.
+ * Currently we only support VXLAN and only one port will actually be
+ * offloaded due to hardware restrictions.
**/
-static void fm10k_add_vxlan_port(struct net_device *dev,
+static void fm10k_udp_tunnel_add(struct net_device *dev,
struct udp_tunnel_info *ti)
{
struct fm10k_intfc *interface = netdev_priv(dev);
- struct fm10k_vxlan_port *vxlan_port;
- if (ti->type != UDP_TUNNEL_TYPE_VXLAN)
- return;
/* only the PF supports configuring tunnels */
if (interface->hw.mac.type != fm10k_mac_pf)
return;
- /* existing ports are pulled out so our new entry is always last */
- fm10k_vxlan_port_for_each(vxlan_port, interface) {
- if ((vxlan_port->port == ti->port) &&
- (vxlan_port->sa_family == ti->sa_family)) {
- list_del(&vxlan_port->list);
- goto insert_tail;
- }
- }
-
- /* allocate memory to track ports */
- vxlan_port = kmalloc(sizeof(*vxlan_port), GFP_ATOMIC);
- if (!vxlan_port)
+ switch (ti->type) {
+ case UDP_TUNNEL_TYPE_VXLAN:
+ fm10k_insert_tunnel_port(&interface->vxlan_port, ti);
+ break;
+ default:
return;
- vxlan_port->port = ti->port;
- vxlan_port->sa_family = ti->sa_family;
-
-insert_tail:
- /* add new port value to list */
- list_add_tail(&vxlan_port->list, &interface->vxlan_port);
+ }
- fm10k_restore_vxlan_port(interface);
+ fm10k_restore_udp_port_info(interface);
}
/**
- * fm10k_del_vxlan_port
+ * fm10k_udp_tunnel_del
* @netdev: network interface device structure
* @ti: Tunnel endpoint information
*
- * This function is called when a new VXLAN interface has freed a port
- * number from the range that is currently in use for VXLAN. The freed
- * port is removed from the list and the new head is used to determine
- * the port number for offloads.
+ * This function is called when a new UDP tunnel port is deleted. The freed
+ * port will be removed from the list, then we reprogram the offloaded port
+ * based on the head of the list.
**/
-static void fm10k_del_vxlan_port(struct net_device *dev,
+static void fm10k_udp_tunnel_del(struct net_device *dev,
struct udp_tunnel_info *ti)
{
struct fm10k_intfc *interface = netdev_priv(dev);
- struct fm10k_vxlan_port *vxlan_port;
+ struct fm10k_udp_port *port = NULL;
- if (ti->type != UDP_TUNNEL_TYPE_VXLAN)
- return;
if (interface->hw.mac.type != fm10k_mac_pf)
return;
- /* find the port in the list and free it */
- fm10k_vxlan_port_for_each(vxlan_port, interface) {
- if ((vxlan_port->port == ti->port) &&
- (vxlan_port->sa_family == ti->sa_family)) {
- list_del(&vxlan_port->list);
- kfree(vxlan_port);
- break;
- }
+ switch (ti->type) {
+ case UDP_TUNNEL_TYPE_VXLAN:
+ port = fm10k_remove_tunnel_port(&interface->vxlan_port, ti);
+ break;
+ default:
+ return;
}
- fm10k_restore_vxlan_port(interface);
+ /* if we did remove a port we need to free its memory */
+ kfree(port);
+
+ fm10k_restore_udp_port_info(interface);
}
/**
@@ -555,7 +573,6 @@ int fm10k_open(struct net_device *netdev)
if (err)
goto err_set_queues;
- /* update VXLAN port configuration */
udp_tunnel_get_rx_info(netdev);
fm10k_up(interface);
@@ -591,7 +608,7 @@ int fm10k_close(struct net_device *netdev)
fm10k_qv_free_irq(interface);
- fm10k_del_vxlan_port_all(interface);
+ fm10k_free_udp_port_info(interface);
fm10k_free_all_tx_resources(interface);
fm10k_free_all_rx_resources(interface);
@@ -1055,7 +1072,7 @@ void fm10k_restore_rx_state(struct fm10k_intfc *interface)
interface->xcast_mode = xcast_mode;
/* Restore tunnel configuration */
- fm10k_restore_vxlan_port(interface);
+ fm10k_restore_udp_port_info(interface);
}
void fm10k_reset_rx_state(struct fm10k_intfc *interface)
@@ -1375,8 +1392,8 @@ static const struct net_device_ops fm10k_netdev_ops = {
.ndo_set_vf_vlan = fm10k_ndo_set_vf_vlan,
.ndo_set_vf_rate = fm10k_ndo_set_vf_bw,
.ndo_get_vf_config = fm10k_ndo_get_vf_config,
- .ndo_udp_tunnel_add = fm10k_add_vxlan_port,
- .ndo_udp_tunnel_del = fm10k_del_vxlan_port,
+ .ndo_udp_tunnel_add = fm10k_udp_tunnel_add,
+ .ndo_udp_tunnel_del = fm10k_udp_tunnel_del,
.ndo_dfwd_add_station = fm10k_dfwd_add_station,
.ndo_dfwd_del_station = fm10k_dfwd_del_station,
#ifdef CONFIG_NET_POLL_CONTROLLER
From 1ad782928f16e7f1b5269ce4358caffe566f44db Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Thu, 23 Jun 2016 13:54:03 -0700
Subject: [PATCH 11/14] fm10k: add support for Rx offloads on one Geneve tunnel
Similar to how we handle VXLAN offload, enable support for a single
Geneve tunnel.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k.h | 1 +
.../net/ethernet/intel/fm10k/fm10k_netdev.c | 30 +++++++++++++++++--
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 3 +-
drivers/net/ethernet/intel/fm10k/fm10k_type.h | 1 +
4 files changed, 31 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k.h b/drivers/net/ethernet/intel/fm10k/fm10k.h
index 209268a147120..67ff01aeb11a6 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k.h
@@ -335,6 +335,7 @@ struct fm10k_intfc {
/* UDP encapsulation port tracking information */
struct list_head vxlan_port;
+ struct list_head geneve_port;
#ifdef CONFIG_DEBUG_FS
struct dentry *dbg_intfc;
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
index 07c5a685f0e94..05629381be6be 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
@@ -387,7 +387,7 @@ static void fm10k_request_glort_range(struct fm10k_intfc *interface)
* fm10k_free_udp_port_info
* @interface: board private structure
*
- * This function frees the entire vxlan_port list
+ * This function frees both geneve_port and vxlan_port structures
**/
static void fm10k_free_udp_port_info(struct fm10k_intfc *interface)
{
@@ -403,6 +403,17 @@ static void fm10k_free_udp_port_info(struct fm10k_intfc *interface)
struct fm10k_udp_port,
list);
}
+
+ /* flush all entries from geneve list */
+ port = list_first_entry_or_null(&interface->geneve_port,
+ struct fm10k_udp_port, list);
+ while (port) {
+ list_del(&port->list);
+ kfree(port);
+ port = list_first_entry_or_null(&interface->vxlan_port,
+ struct fm10k_udp_port,
+ list);
+ }
}
/**
@@ -427,6 +438,13 @@ static void fm10k_restore_udp_port_info(struct fm10k_intfc *interface)
fm10k_write_reg(hw, FM10K_TUNNEL_CFG,
(port ? ntohs(port->port) : 0) |
(ETH_P_TEB << FM10K_TUNNEL_CFG_NVGRE_SHIFT));
+
+ port = list_first_entry_or_null(&interface->geneve_port,
+ struct fm10k_udp_port, list);
+
+ /* restore Geneve tunnel configuration register */
+ fm10k_write_reg(hw, FM10K_TUNNEL_CFG_GENEVE,
+ (port ? ntohs(port->port) : 0));
}
static struct fm10k_udp_port *
@@ -472,8 +490,8 @@ static void fm10k_insert_tunnel_port(struct list_head *ports,
* @ti: Tunnel endpoint information
*
* This function is called when a new UDP tunnel port has been added.
- * Currently we only support VXLAN and only one port will actually be
- * offloaded due to hardware restrictions.
+ * Due to hardware restrictions, only one port per type can be offloaded at
+ * once.
**/
static void fm10k_udp_tunnel_add(struct net_device *dev,
struct udp_tunnel_info *ti)
@@ -488,6 +506,9 @@ static void fm10k_udp_tunnel_add(struct net_device *dev,
case UDP_TUNNEL_TYPE_VXLAN:
fm10k_insert_tunnel_port(&interface->vxlan_port, ti);
break;
+ case UDP_TUNNEL_TYPE_GENEVE:
+ fm10k_insert_tunnel_port(&interface->geneve_port, ti);
+ break;
default:
return;
}
@@ -517,6 +538,9 @@ static void fm10k_udp_tunnel_del(struct net_device *dev,
case UDP_TUNNEL_TYPE_VXLAN:
port = fm10k_remove_tunnel_port(&interface->vxlan_port, ti);
break;
+ case UDP_TUNNEL_TYPE_GENEVE:
+ port = fm10k_remove_tunnel_port(&interface->geneve_port, ti);
+ break;
default:
return;
}
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index 860fe04f4c72b..0a20ca1685354 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -1835,8 +1835,9 @@ static int fm10k_sw_init(struct fm10k_intfc *interface,
interface->tx_itr = FM10K_TX_ITR_DEFAULT;
interface->rx_itr = FM10K_ITR_ADAPTIVE | FM10K_RX_ITR_DEFAULT;
- /* initialize vxlan_port list */
+ /* initialize udp port lists */
INIT_LIST_HEAD(&interface->vxlan_port);
+ INIT_LIST_HEAD(&interface->geneve_port);
netdev_rss_key_fill(rss_key, sizeof(rss_key));
memcpy(interface->rssrk, rss_key, sizeof(rss_key));
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_type.h b/drivers/net/ethernet/intel/fm10k/fm10k_type.h
index f4e75c498287c..6bb16c13d9d6c 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_type.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_type.h
@@ -154,6 +154,7 @@ struct fm10k_hw;
#define FM10K_DGLORTDEC_INNERRSS_ENABLE 0x08000000
#define FM10K_TUNNEL_CFG 0x0040
#define FM10K_TUNNEL_CFG_NVGRE_SHIFT 16
+#define FM10K_TUNNEL_CFG_GENEVE 0x0041
#define FM10K_SWPRI_MAP(_n) ((_n) + 0x0050)
#define FM10K_SWPRI_MAX 16
#define FM10K_RSSRK(_n, _m) (((_n) * 0x10) + (_m) + 0x0800)
From 9717c7721302d217fb73f59eb3ab6314ffff54aa Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Fri, 22 Jul 2016 16:00:37 -0700
Subject: [PATCH 12/14] fm10k: remove unnecessary extra parenthesis around
((~value))
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index 0a20ca1685354..d7f1ecad93781 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -133,7 +133,7 @@ static void fm10k_detach_subtask(struct fm10k_intfc *interface)
/* check the real address space to see if we've recovered */
hw_addr = READ_ONCE(interface->uc_addr);
value = readl(hw_addr);
- if ((~value)) {
+ if (~value) {
interface->hw.hw_addr = interface->uc_addr;
netif_device_attach(netdev);
interface->flags |= FM10K_FLAG_RESET_REQUESTED;
From c689eff124cb231d91777a447fa05b30939da7b1 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Wed, 3 Aug 2016 15:05:27 -0700
Subject: [PATCH 13/14] fm10k: don't clear the RXQCTL register when enabling or
disabling queues
Ensure that other bits in the RXQCTL register do not get cleared. This
ensures that bits related to queue ownership are maintained.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
index d7f1ecad93781..b1a2f8437d596 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pci.c
@@ -734,15 +734,15 @@ static void fm10k_configure_rx_ring(struct fm10k_intfc *interface,
u64 rdba = ring->dma;
struct fm10k_hw *hw = &interface->hw;
u32 size = ring->count * sizeof(union fm10k_rx_desc);
- u32 rxqctl = FM10K_RXQCTL_ENABLE | FM10K_RXQCTL_PF;
- u32 rxdctl = FM10K_RXDCTL_WRITE_BACK_MIN_DELAY;
+ u32 rxqctl, rxdctl = FM10K_RXDCTL_WRITE_BACK_MIN_DELAY;
u32 srrctl = FM10K_SRRCTL_BUFFER_CHAINING_EN;
u32 rxint = FM10K_INT_MAP_DISABLE;
u8 rx_pause = interface->rx_pause;
u8 reg_idx = ring->reg_idx;
/* disable queue to avoid issues while updating state */
- fm10k_write_reg(hw, FM10K_RXQCTL(reg_idx), 0);
+ rxqctl = fm10k_read_reg(hw, FM10K_RXQCTL(reg_idx));
+ rxqctl &= ~FM10K_RXQCTL_ENABLE;
fm10k_write_flush(hw);
/* possible poll here to verify ring resources have been cleaned */
@@ -797,6 +797,8 @@ static void fm10k_configure_rx_ring(struct fm10k_intfc *interface,
fm10k_write_reg(hw, FM10K_RXINT(reg_idx), rxint);
/* enable queue */
+ rxqctl = fm10k_read_reg(hw, FM10K_RXQCTL(reg_idx));
+ rxqctl |= FM10K_RXQCTL_ENABLE;
fm10k_write_reg(hw, FM10K_RXQCTL(reg_idx), rxqctl);
/* place buffers on ring for receive data */
From 325782a173d1858bb67a827905e264cd128241d0 Mon Sep 17 00:00:00 2001
From: Jacob Keller <jacob.e.keller@intel.com>
Date: Mon, 8 Aug 2016 17:08:19 -0700
Subject: [PATCH 14/14] fm10k: don't re-map queues when a mailbox message
suffices
When the PF assigns a new MAC address to a VF it uses the base address
registers to store the MAC address. This allows a VF which loads after
this setup the ability to get the initial address without having to wait
for a mailbox message. Unfortunately to do this, the PF must take queue
ownership away from the VF, which can cause fault errors when there is
already an active VF driver.
This queue ownership assignment causes race condition between the PF and
the VF such that potentially a VF can cause FUM fault errors due to
normal PF/VF driver behavior.
It is not safe to simply allow the PF to write the base address
registers without taking queue ownership back as the PF must also
disable the queues, and this would impact active VF use. The current
code is safe because the queue ownership will prevent the VF from
actually writing but does trigger the FUM fault.
We can do better by simply avoiding the register write process when
a mailbox message suffices. If the message can be sent over the mailbox,
then we will not perform the queue ownership assignment and we won't
update the base address to be the same as the MAC address.
We do still have to write the TXQCTL registers in order to update the
VID of the queue. This is necessary because the TXQCTL register is
read-only from the VF, and thus the VF cannot do this for itself. This
register does not need to wait for the Tx queue to be disabled and is
safe for the PF to write during normal VF operation, so we move this
write to the top of the function above the mailbox message. Without
this, the TXQCTL register would be misconfigured and cause the VF to Tx
hang.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/fm10k/fm10k_pf.c | 46 +++++++++++++--------
1 file changed, 29 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_pf.c b/drivers/net/ethernet/intel/fm10k/fm10k_pf.c
index 682299dd0ce41..23fb319fd2a04 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_pf.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_pf.c
@@ -867,10 +867,6 @@ static s32 fm10k_iov_assign_default_mac_vlan_pf(struct fm10k_hw *hw,
vf_q_idx = fm10k_vf_queue_index(hw, vf_idx);
qmap_idx = qmap_stride * vf_idx;
- /* MAP Tx queue back to 0 temporarily, and disable it */
- fm10k_write_reg(hw, FM10K_TQMAP(qmap_idx), 0);
- fm10k_write_reg(hw, FM10K_TXDCTL(vf_q_idx), 0);
-
/* Determine correct default VLAN ID. The FM10K_VLAN_OVERRIDE bit is
* used here to indicate to the VF that it will not have privilege to
* write VLAN_TABLE. All policy is enforced on the PF but this allows
@@ -886,9 +882,35 @@ static s32 fm10k_iov_assign_default_mac_vlan_pf(struct fm10k_hw *hw,
fm10k_tlv_attr_put_mac_vlan(msg, FM10K_MAC_VLAN_MSG_DEFAULT_MAC,
vf_info->mac, vf_vid);
- /* load onto outgoing mailbox, ignore any errors on enqueue */
- if (vf_info->mbx.ops.enqueue_tx)
- vf_info->mbx.ops.enqueue_tx(hw, &vf_info->mbx, msg);
+ /* Configure Queue control register with new VLAN ID. The TXQCTL
+ * register is RO from the VF, so the PF must do this even in the
+ * case of notifying the VF of a new VID via the mailbox.
+ */
+ txqctl = ((u32)vf_vid << FM10K_TXQCTL_VID_SHIFT) &
+ FM10K_TXQCTL_VID_MASK;
+ txqctl |= (vf_idx << FM10K_TXQCTL_TC_SHIFT) |
+ FM10K_TXQCTL_VF | vf_idx;
+
+ for (i = 0; i < queues_per_pool; i++)
+ fm10k_write_reg(hw, FM10K_TXQCTL(vf_q_idx + i), txqctl);
+
+ /* try loading a message onto outgoing mailbox first */
+ if (vf_info->mbx.ops.enqueue_tx) {
+ err = vf_info->mbx.ops.enqueue_tx(hw, &vf_info->mbx, msg);
+ if (err != FM10K_MBX_ERR_NO_MBX)
+ return err;
+ err = 0;
+ }
+
+ /* If we aren't connected to a mailbox, this is most likely because
+ * the VF driver is not running. It should thus be safe to re-map
+ * queues and use the registers to pass the MAC address so that the VF
+ * driver gets correct information during its initialization.
+ */
+
+ /* MAP Tx queue back to 0 temporarily, and disable it */
+ fm10k_write_reg(hw, FM10K_TQMAP(qmap_idx), 0);
+ fm10k_write_reg(hw, FM10K_TXDCTL(vf_q_idx), 0);
/* verify ring has disabled before modifying base address registers */
txdctl = fm10k_read_reg(hw, FM10K_TXDCTL(vf_q_idx));
@@ -927,16 +949,6 @@ static s32 fm10k_iov_assign_default_mac_vlan_pf(struct fm10k_hw *hw,
FM10K_TDLEN_ITR_SCALE_SHIFT);
err_out:
- /* configure Queue control register */
- txqctl = ((u32)vf_vid << FM10K_TXQCTL_VID_SHIFT) &
- FM10K_TXQCTL_VID_MASK;
- txqctl |= (vf_idx << FM10K_TXQCTL_TC_SHIFT) |
- FM10K_TXQCTL_VF | vf_idx;
-
- /* assign VLAN ID */
- for (i = 0; i < queues_per_pool; i++)
- fm10k_write_reg(hw, FM10K_TXQCTL(vf_q_idx + i), txqctl);
-
/* restore the queue back to VF ownership */
fm10k_write_reg(hw, FM10K_TQMAP(qmap_idx), vf_q_idx);
return err;