[PATCH 0/2] be2net: patch set

[PATCH 0/2] be2net: patch set

[Sathya Perla]

Patch 1/2 is a v2 of a patch that was submitted earlier (as a part of a
different patch-set). v2 incorporates a suggestion given by David Laight
for how long to poll for pending TX completions while disabling a device.

Patch 2/2 fixes a crash in be_remove()->be_close()
path after be2net has aborted an EEH error recovery
due to a permanant failure.

Kalesh AP (1):
  be2net: Fix invocation of be_close() after be_clear()

Vasundhara Volam (1):
  be2net: Fix to reap TX compls till HW doesn't respond for some time

 drivers/net/ethernet/emulex/benet/be.h      |    1 +
 drivers/net/ethernet/emulex/benet/be_main.c |   17 +++++++++++++----
 2 files changed, 14 insertions(+), 4 deletions(-)

[PATCH 1/2] be2net: Fix to reap TX compls till HW doesn't respond for some time

[Sathya Perla]

From: Vasundhara Volam <vasundhara.volam@emulex.com>

be_close() currently waits for a max of 200ms to receive all pending
TX compls. This timeout value was roughly calculated based on 10G
transmission speeds and the TX queue depth. This timeout may not be
enough when the link is operating at lower speeds or in multi-channel/SR-IOV
configs with TX-rate limiting setting.

It is hard to calculate a "proper timeout value" that works in all
configurations.  This patch solves this problem by continuing to reap
TX completions till the HW is completely silent for a period of 10ms or
a HW error is detected.

v2: implements the new scheme (as suggested by David Laight) instead of
just waiting longer than 200ms for reaping all completions.

Signed-off-by: Vasundhara Volam <vasundhara.volam@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
 drivers/net/ethernet/emulex/benet/be_main.c |    9 +++++----
 1 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 3e6df47..80f754d 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -2033,11 +2033,13 @@ static void be_tx_compl_clean(struct be_adapter *adapter)
 	bool dummy_wrb;
 	int i, pending_txqs;
 
-	/* Wait for a max of 200ms for all the tx-completions to arrive. */
+	/* Stop polling for compls when HW has been silent for 10ms */
 	do {
 		pending_txqs = adapter->num_tx_qs;
 
 		for_all_tx_queues(adapter, txo, i) {
+			cmpl = 0;
+			num_wrbs = 0;
 			txq = &txo->q;
 			while ((txcp = be_tx_compl_get(&txo->cq))) {
 				end_idx =
@@ -2050,14 +2052,13 @@ static void be_tx_compl_clean(struct be_adapter *adapter)
 			if (cmpl) {
 				be_cq_notify(adapter, txo->cq.id, false, cmpl);
 				atomic_sub(num_wrbs, &txq->used);
-				cmpl = 0;
-				num_wrbs = 0;
+				timeo = 0;
 			}
 			if (atomic_read(&txq->used) == 0)
 				pending_txqs--;
 		}
 
-		if (pending_txqs == 0 || ++timeo > 200)
+		if (pending_txqs == 0 || ++timeo > 10 || be_hw_error(adapter))
 			break;
 
 		mdelay(1);
-- 
1.7.1

[PATCH 2/2] be2net: Fix invocation of be_close() after be_clear()

[Sathya Perla]

From: Kalesh AP <kalesh.purayil@emulex.com>

In the EEH error recovery path, when a permanent failure occurs,
we clean up adapter structure (i.e. destroy queues etc) by calling
be_clear() and return PCI_ERS_RESULT_DISCONNECT.
After this the stack tries to remove device from bus and calls
be_remove() which invokes netdev_unregister()->be_close().
be_close() operating on destroyed queues results in a
NULL dereference.

This patch fixes this problem by introducing a flag to keep track
of the setup state.

Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
---
 drivers/net/ethernet/emulex/benet/be.h      |    1 +
 drivers/net/ethernet/emulex/benet/be_main.c |    8 ++++++++
 2 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 8ccaa25..97db5a7 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -374,6 +374,7 @@ enum vf_state {
 #define BE_FLAGS_NAPI_ENABLED			(1 << 9)
 #define BE_FLAGS_QNQ_ASYNC_EVT_RCVD		(1 << 11)
 #define BE_FLAGS_VXLAN_OFFLOADS			(1 << 12)
+#define BE_FLAGS_SETUP_DONE			(1 << 13)
 
 #define BE_UC_PMAC_COUNT			30
 #define BE_VF_UC_PMAC_COUNT			2
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 80f754d..a186454 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -2726,6 +2726,12 @@ static int be_close(struct net_device *netdev)
 	struct be_eq_obj *eqo;
 	int i;
 
+	/* This protection is needed as be_close() may be called even when the
+	 * adapter is in cleared state (after eeh perm failure)
+	 */
+	if (!(adapter->flags & BE_FLAGS_SETUP_DONE))
+		return 0;
+
 	be_roce_dev_close(adapter);
 
 	if (adapter->flags & BE_FLAGS_NAPI_ENABLED) {
@@ -3056,6 +3062,7 @@ static int be_clear(struct be_adapter *adapter)
 	be_clear_queues(adapter);
 
 	be_msix_disable(adapter);
+	adapter->flags &= ~BE_FLAGS_SETUP_DONE;
 	return 0;
 }
 
@@ -3560,6 +3567,7 @@ static int be_setup(struct be_adapter *adapter)
 		adapter->phy.fc_autoneg = 1;
 
 	be_schedule_worker(adapter);
+	adapter->flags |= BE_FLAGS_SETUP_DONE;
 	return 0;
 err:
 	be_clear(adapter);
-- 
1.7.1

Re: [PATCH 0/2] be2net: patch set

[David Miller]

From: Sathya Perla <sathya.perla@emulex.com>
Date: Mon, 14 Apr 2014 16:12:39 +0530

> Patch 1/2 is a v2 of a patch that was submitted earlier (as a part of a
> different patch-set). v2 incorporates a suggestion given by David Laight
> for how long to poll for pending TX completions while disabling a device.
> 
> Patch 2/2 fixes a crash in be_remove()->be_close()
> path after be2net has aborted an EEH error recovery
> due to a permanant failure.

Series applied, thank you.