[PATCH 0/5] be2net: patch set

Patch 1 fixes a FW image compatibility check in the driver that
prevents certain FW images from being flashed on BE3 (not BE3-R)
adapters.

Patch 2 fixes a spin_lock not being released in a failure case in
be_cmd_notify_wait().

Patch 3 includes a workaround to pad packets that are only 32b long or less
to be applicabe to BE3 too. This workaround was currently applied only to
Skyhawk and Lancer chips. Such packets are causing BE3's TX path to stall
on a SR-IOV config.

Patch 4 fixes the be_cmd_get_profile_config() routine to set the pf_num
field in the cmd request. The FW requires this field to be set for it to
return the specific function's descriptors. If not set, the FW returns
the descriptors of all the functions on the device. If the first descriptor
is not what is being queried for, the driver will read wrong data.
This patch fixes this issue by using the GET_CNTL_ATTRIB cmd to query the
real pci_func_num of a function and then uses it in the GET_PROFILE_CONFIG
cmd.

Patch 5 completes an earlier fix that removed the vlan promisc capability
for VFs. The earlier fix did not update the removal of this capability from
the profile descriptor of the VF. This causes the VF driver to request this
capability when it tries to create it's interface at probe time. This could
potentailly cause the VF probe to fail if the FW enforces strict checking of
the flags based on what was provisoned by the PF.  This strict checking is
not being done by FW currently but will be fixed in a future version. This
patch fixes this issue by updating the VF's profile descriptor so that they
match the interface capability flags provisioned by the PF.

Pls consider adding these patches to the net tree. Thanks!

Kalesh AP (2):
  be2net: fix BE3-R FW download compatibility check
  be2net: remove vlan promisc capability from VF's profile descriptors

Somnath Kotur (1):
  be2net: set pci_func_num while issuing GET_PROFILE_CONFIG cmd

Suresh Reddy (2):
  be2net: release mcc-lock in a failure case in be_cmd_notify_wait()
  be2net: pad skb to meet minimum TX pkt size in BE3

 drivers/net/ethernet/emulex/benet/be.h      |  1 +
 drivers/net/ethernet/emulex/benet/be_cmds.c | 31 ++++++++++++++++++++++-------
 drivers/net/ethernet/emulex/benet/be_cmds.h | 10 ++++++++--
 drivers/net/ethernet/emulex/benet/be_main.c | 28 +++++++++++++++++---------
 4 files changed, 52 insertions(+), 18 deletions(-)

-- 
2.4.1

From: Kalesh AP <redacted>

In the BE3 FW image, unlike Skyhawk's, the "asic_type_rev" field doesn't
track the asic_rev of chip it is compatible with. When asic_type_rev
is 0 the image is compatible only with pre-BE3-R chips (asic_rev < 0x10).
Fix the current compatibility check to take care of this.
We hit this issue when we try to flash old BE3 images (used prior to the
release of BE3-R) on pre-BE3-R adapters.

Fixes: a6e6ff6eee12f3e ("be2net: simplify UFI compatibility checking")
Signed-off-by: Kalesh AP <redacted>
Signed-off-by: Sathya Perla <redacted>
---
 drivers/net/ethernet/emulex/benet/be_main.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 7bf51a1..86eed47 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c

@@ -4999,7 +4999,15 @@ static bool be_check_ufi_compatibility(struct be_adapter *adapter,
 		return false;
 	}
 
-	return (fhdr->asic_type_rev >= adapter->asic_rev);
+	/* In BE3 FW images the "asic_type_rev" field doesn't track the
+	 * asic_rev of the chips it is compatible with.
+	 * When asic_type_rev is 0 the image is compatible only with
+	 * pre-BE3-R chips (asic_rev < 0x10)
+	 */
+	if (BEx_chip(adapter) && fhdr->asic_type_rev == 0)
+		return adapter->asic_rev < 0x10;
+	else
+		return (fhdr->asic_type_rev >= adapter->asic_rev);
 }
 
 static int be_fw_download(struct be_adapter *adapter, const struct firmware* fw)

-- 
2.4.1

From: Suresh Reddy <redacted>

The mcc/mbox lock is not being released when be_cmd_copy() returns
an error.

Signed-off-by: Suresh Reddy <redacted>
Signed-off-by: Sathya Perla <redacted>
---
 drivers/net/ethernet/emulex/benet/be_cmds.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index eb32391..9dc5ce1 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c

@@ -851,8 +851,10 @@ static int be_cmd_notify_wait(struct be_adapter *adapter,
 		return status;
 
 	dest_wrb = be_cmd_copy(adapter, wrb);
-	if (!dest_wrb)
-		return -EBUSY;
+	if (!dest_wrb) {
+		status = -EBUSY;
+		goto unlock;
+	}
 
 	if (use_mcc(adapter))
 		status = be_mcc_notify_wait(adapter);

@@ -862,6 +864,7 @@ static int be_cmd_notify_wait(struct be_adapter *adapter,
 	if (!status)
 		memcpy(wrb, dest_wrb, sizeof(*wrb));
 
+unlock:
 	be_cmd_unlock(adapter);
 	return status;
 }

-- 
2.4.1

From: Suresh Reddy <redacted>

On BE3 chips in SRIOV configs, the TX path stalls when a packet less
than 32B is received from the host. A workaround to pad such packets
already exists for the Skyhawk and Lancer chips. Use the same workaround
for BE3 chips too.

Signed-off-by: Suresh Reddy <redacted>
Signed-off-by: Sathya Perla <redacted>
---
 drivers/net/ethernet/emulex/benet/be_main.c | 9 +++++----
 1 file 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 86eed47..821e014 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c

@@ -1123,11 +1123,12 @@ static struct sk_buff *be_xmit_workarounds(struct be_adapter *adapter,
 					   struct sk_buff *skb,
 					   struct be_wrb_params *wrb_params)
 {
-	/* Lancer, SH-R ASICs have a bug wherein Packets that are 32 bytes or
-	 * less may cause a transmit stall on that port. So the work-around is
-	 * to pad short packets (<= 32 bytes) to a 36-byte length.
+	/* Lancer, SH and BE3 in SRIOV mode have a bug wherein
+	 * packets that are 32b or less may cause a transmit stall
+	 * on that port. The workaround is to pad such packets
+	 * (len <= 32 bytes) to a minimum length of 36b.
 	 */
-	if (unlikely(!BEx_chip(adapter) && skb->len <= 32)) {
+	if (skb->len <= 32) {
 		if (skb_put_padto(skb, 36))
 			return NULL;
 	}

-- 
2.4.1

From: Somnath Kotur <redacted>

The FW requires the pf_num field in the cmd hdr to be set for it to return
the specific function's descriptors in the GET_PROFILE_CONFIG cmd. If not
set, the FW returns the descriptors of all the functions on the device.
If the first descriptor is not what is being queried for, the driver will
read wrong data. This patch fixes this issue by using the GET_CNTL_ATTRIB
cmd to query the real pci_func_num of a function and then uses it in the
GET_PROFILE_CONFIG cmd.

Signed-off-by: Somnath Kotur <redacted>
Signed-off-by: Sathya Perla <redacted>
---
 drivers/net/ethernet/emulex/benet/be.h      |  1 +
 drivers/net/ethernet/emulex/benet/be_cmds.c | 14 +++++++++++---
 drivers/net/ethernet/emulex/benet/be_cmds.h | 10 ++++++++--
 drivers/net/ethernet/emulex/benet/be_main.c |  9 +++++----
 4 files changed, 25 insertions(+), 9 deletions(-)

diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 8215409..d463563 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h

@@ -592,6 +592,7 @@ struct be_adapter {
 	int be_get_temp_freq;
 	struct be_hwmon hwmon_info;
 	u8 pf_number;
+	u8 pci_func_num;
 	struct rss_info rss_info;
 	/* Filters for packets that need to be sent to BMC */
 	u32 bmc_filt_mask;

diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index 9dc5ce1..790284d 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c

@@ -2890,6 +2890,7 @@ int be_cmd_get_cntl_attributes(struct be_adapter *adapter)
 	if (!status) {
 		attribs = attribs_cmd.va + sizeof(struct be_cmd_resp_hdr);
 		adapter->hba_port_num = attribs->hba_attribs.phy_port;
+		adapter->pci_func_num = attribs->pci_func_num;
 		serial_num = attribs->hba_attribs.controller_serial_number;
 		for (i = 0; i < CNTL_SERIAL_NUM_WORDS; i++)
 			adapter->serial_num[i] = le32_to_cpu(serial_num[i]) &

@@ -3712,7 +3713,6 @@ int be_cmd_get_func_config(struct be_adapter *adapter, struct be_resources *res)
 			status = -EINVAL;
 			goto err;
 		}
-
 		adapter->pf_number = desc->pf_num;
 		be_copy_nic_desc(res, desc);
 	}

@@ -3724,7 +3724,10 @@ err:
 	return status;
 }
 
-/* Will use MBOX only if MCCQ has not been created */
+/* Will use MBOX only if MCCQ has not been created
+ * non-zero domain => a PF is querying this on behalf of a VF
+ * zero domain => a PF or a VF is querying this for itself
+ */
 int be_cmd_get_profile_config(struct be_adapter *adapter,
 			      struct be_resources *res, u8 query, u8 domain)
 {

@@ -3751,10 +3754,15 @@ int be_cmd_get_profile_config(struct be_adapter *adapter,
 			       OPCODE_COMMON_GET_PROFILE_CONFIG,
 			       cmd.size, &wrb, &cmd);
 
-	req->hdr.domain = domain;
 	if (!lancer_chip(adapter))
 		req->hdr.version = 1;
 	req->type = ACTIVE_PROFILE_TYPE;
+	/* When a function is querying profile information relating to
+	 * itself hdr.pf_number must be set to it's pci_func_num + 1
+	 */
+	req->hdr.domain = domain;
+	if (domain == 0)
+		req->hdr.pf_num = adapter->pci_func_num + 1;
 
 	/* When QUERY_MODIFIABLE_FIELDS_TYPE bit is set, cmd returns the
 	 * descriptors with all bits set to "1" for the fields which can be

diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.h b/drivers/net/ethernet/emulex/benet/be_cmds.h
index 7d178bd..91155ea 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.h
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.h

@@ -289,7 +289,9 @@ struct be_cmd_req_hdr {
 	u32 timeout;		/* dword 1 */
 	u32 request_length;	/* dword 2 */
 	u8 version;		/* dword 3 */
-	u8 rsvd[3];		/* dword 3 */
+	u8 rsvd1;		/* dword 3 */
+	u8 pf_num;		/* dword 3 */
+	u8 rsvd2;		/* dword 3 */
 };
 
 #define RESP_HDR_INFO_OPCODE_SHIFT	0	/* bits 0 - 7 */

@@ -1652,7 +1654,11 @@ struct mgmt_hba_attribs {
 
 struct mgmt_controller_attrib {
 	struct mgmt_hba_attribs hba_attribs;
-	u32 rsvd0[10];
+	u32 rsvd0[2];
+	u16 rsvd1;
+	u8 pci_func_num;
+	u8 rsvd2;
+	u32 rsvd3[7];
 } __packed;
 
 struct be_cmd_req_cntl_attribs {

diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 821e014..eb48a97 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c

@@ -4206,10 +4206,6 @@ static int be_get_config(struct be_adapter *adapter)
 	int status, level;
 	u16 profile_id;
 
-	status = be_cmd_get_cntl_attributes(adapter);
-	if (status)
-		return status;
-
 	status = be_cmd_query_fw_cfg(adapter);
 	if (status)
 		return status;

@@ -4408,6 +4404,11 @@ static int be_setup(struct be_adapter *adapter)
 	if (!lancer_chip(adapter))
 		be_cmd_req_native_mode(adapter);
 
+	/* Need to invoke this cmd first to get the PCI Function Number */
+	status = be_cmd_get_cntl_attributes(adapter);
+	if (status)
+		return status;
+
 	if (!BE2_chip(adapter) && be_physfn(adapter))
 		be_alloc_sriov_res(adapter);
 

-- 
2.4.1

From: Kalesh AP <redacted>

The commit 435452aa8847 ("Prevent VFs from enabling VLAN promiscuous mode")
fixed the PF driver to not include the VLAN promisc capability while
provisioning the interface for a VF. But the fix did not remove this
capability from the profile descriptor of the VF. This causes the VF
driver to request this capability when it tries to create it's interface
at probe time.  This could potentailly cause the VF probe to fail if the
FW enforces strict checking of the flags based on what was provisoned
by the PF.  This strict checking is not being done by FW currently but
will be fixed in a future version. This patch fixes this issue by updating
the VF's profile descriptor so that they match the interface capability
flags provisioned by the PF.

Fixes: 435452aa8847 ("Prevent VFs from enabling VLAN promiscuous mode")
Signed-off-by: Kalesh AP <redacted>
Signed-off-by: Sathya Perla <redacted>
---
 drivers/net/ethernet/emulex/benet/be_cmds.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index 790284d..1795c93 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c

@@ -1987,6 +1987,8 @@ int be_cmd_rx_filter(struct be_adapter *adapter, u32 flags, u32 value)
 			 be_if_cap_flags(adapter));
 	}
 	flags &= be_if_cap_flags(adapter);
+	if (!flags)
+		return -ENOTSUPP;
 
 	return __be_cmd_rx_filter(adapter, flags, value);
 }

@@ -3932,12 +3934,16 @@ static void be_fill_vf_res_template(struct be_adapter *adapter,
 			vf_if_cap_flags &= ~(BE_IF_FLAGS_RSS |
 					     BE_IF_FLAGS_DEFQ_RSS);
 		}
-
-		nic_vft->cap_flags = cpu_to_le32(vf_if_cap_flags);
 	} else {
 		num_vf_qs = 1;
 	}
 
+	if (res_mod.vf_if_cap_flags & BE_IF_FLAGS_VLAN_PROMISCUOUS) {
+		nic_vft->flags |= BIT(IF_CAPS_FLAGS_VALID_SHIFT);
+		vf_if_cap_flags &= ~BE_IF_FLAGS_VLAN_PROMISCUOUS;
+	}
+
+	nic_vft->cap_flags = cpu_to_le32(vf_if_cap_flags);
 	nic_vft->rq_count = cpu_to_le16(num_vf_qs);
 	nic_vft->txq_count = cpu_to_le16(num_vf_qs);
 	nic_vft->rssq_count = cpu_to_le16(num_vf_qs);

-- 
2.4.1

From: Sathya Perla <redacted>
Date: Mon, 12 Oct 2015 03:47:16 -0400

Patch 1 fixes a FW image compatibility check in the driver that
prevents certain FW images from being flashed on BE3 (not BE3-R)
adapters.

Patch 2 fixes a spin_lock not being released in a failure case in
be_cmd_notify_wait().

Patch 3 includes a workaround to pad packets that are only 32b long or less
to be applicabe to BE3 too. This workaround was currently applied only to
Skyhawk and Lancer chips. Such packets are causing BE3's TX path to stall
on a SR-IOV config.

Patch 4 fixes the be_cmd_get_profile_config() routine to set the pf_num
field in the cmd request. The FW requires this field to be set for it to
return the specific function's descriptors. If not set, the FW returns
the descriptors of all the functions on the device. If the first descriptor
is not what is being queried for, the driver will read wrong data.
This patch fixes this issue by using the GET_CNTL_ATTRIB cmd to query the
real pci_func_num of a function and then uses it in the GET_PROFILE_CONFIG
cmd.

Patch 5 completes an earlier fix that removed the vlan promisc capability
for VFs. The earlier fix did not update the removal of this capability from
the profile descriptor of the VF. This causes the VF driver to request this
capability when it tries to create it's interface at probe time. This could
potentailly cause the VF probe to fail if the FW enforces strict checking of
the flags based on what was provisoned by the PF.  This strict checking is
not being done by FW currently but will be fixed in a future version. This
patch fixes this issue by updating the VF's profile descriptor so that they
match the interface capability flags provisioned by the PF.

Pls consider adding these patches to the net tree. Thanks!

Series applied, thanks.