[PATCH net-next 0/3] dsa: b53/sf2

[PATCH net-next 0/3] dsa: b53/sf2

[Andrew Lunn]

Fixup most of the C=1 W=1 warnings in these drivers.

Andrew Lunn (3):
  net: dsa: b53: Fixup endianness warnings
  dsa: bcm_sf2: Initialize __be16 with a __be16 value
  dsa: bmc_sf2: Pass GENMASK() signed bits

 drivers/net/dsa/b53/b53_spi.c | 26 ++++++++++++++++++--------
 drivers/net/dsa/bcm_sf2_cfp.c |  8 ++++----
 2 files changed, 22 insertions(+), 12 deletions(-)

-- 
2.27.0.rc2

[PATCH net-next 1/3] net: dsa: b53: Fixup endianness warnings

[Andrew Lunn]

leX_to_cpu() expects to be passed an __leX type.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/b53/b53_spi.c | 26 ++++++++++++++++++--------
 1 file changed, 18 insertions(+), 8 deletions(-)

diff --git a/drivers/net/dsa/b53/b53_spi.c b/drivers/net/dsa/b53/b53_spi.c
index f89f5308a99b..7abec8dab8ba 100644
--- a/drivers/net/dsa/b53/b53_spi.c
+++ b/drivers/net/dsa/b53/b53_spi.c
@@ -145,42 +145,52 @@ static int b53_spi_read8(struct b53_device *dev, u8 page, u8 reg, u8 *val)
 
 static int b53_spi_read16(struct b53_device *dev, u8 page, u8 reg, u16 *val)
 {
-	int ret = b53_spi_read(dev, page, reg, (u8 *)val, 2);
+	__le16 value;
+	int ret;
+
+	ret = b53_spi_read(dev, page, reg, (u8 *)&value, 2);
 
 	if (!ret)
-		*val = le16_to_cpu(*val);
+		*val = le16_to_cpu(value);
 
 	return ret;
 }
 
 static int b53_spi_read32(struct b53_device *dev, u8 page, u8 reg, u32 *val)
 {
-	int ret = b53_spi_read(dev, page, reg, (u8 *)val, 4);
+	__le32 value;
+	int ret;
+
+	ret = b53_spi_read(dev, page, reg, (u8 *)&value, 4);
 
 	if (!ret)
-		*val = le32_to_cpu(*val);
+		*val = le32_to_cpu(value);
 
 	return ret;
 }
 
 static int b53_spi_read48(struct b53_device *dev, u8 page, u8 reg, u64 *val)
 {
+	__le64 value;
 	int ret;
 
 	*val = 0;
-	ret = b53_spi_read(dev, page, reg, (u8 *)val, 6);
+	ret = b53_spi_read(dev, page, reg, (u8 *)&value, 6);
 	if (!ret)
-		*val = le64_to_cpu(*val);
+		*val = le64_to_cpu(value);
 
 	return ret;
 }
 
 static int b53_spi_read64(struct b53_device *dev, u8 page, u8 reg, u64 *val)
 {
-	int ret = b53_spi_read(dev, page, reg, (u8 *)val, 8);
+	__le64 value;
+	int ret;
+
+	ret = b53_spi_read(dev, page, reg, (u8 *)&value, 8);
 
 	if (!ret)
-		*val = le64_to_cpu(*val);
+		*val = le64_to_cpu(value);
 
 	return ret;
 }
-- 
2.27.0.rc2

[PATCH net-next 2/3] dsa: bcm_sf2: Initialize __be16 with a __be16 value

[Andrew Lunn]

A __be16 variable should be initialised with a __be16 value.  So add a
htons(). In this case it is pointless, given the value being assigned
is 0xffff, but it stops sparse from warnings.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/bcm_sf2_cfp.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dsa/bcm_sf2_cfp.c b/drivers/net/dsa/bcm_sf2_cfp.c
index f707edc641cf..b54339477853 100644
--- a/drivers/net/dsa/bcm_sf2_cfp.c
+++ b/drivers/net/dsa/bcm_sf2_cfp.c
@@ -348,8 +348,8 @@ static int bcm_sf2_cfp_ipv4_rule_set(struct bcm_sf2_priv *priv, int port,
 				     unsigned int queue_num,
 				     struct ethtool_rx_flow_spec *fs)
 {
+	__be16 vlan_tci = 0, vlan_m_tci = htons(0xffff);
 	struct ethtool_rx_flow_spec_input input = {};
-	__be16 vlan_tci = 0 , vlan_m_tci = 0xffff;
 	const struct cfp_udf_layout *layout;
 	unsigned int slice_num, rule_index;
 	struct ethtool_rx_flow_rule *flow;
@@ -629,8 +629,8 @@ static int bcm_sf2_cfp_ipv6_rule_set(struct bcm_sf2_priv *priv, int port,
 				     unsigned int queue_num,
 				     struct ethtool_rx_flow_spec *fs)
 {
+	__be16 vlan_tci = 0, vlan_m_tci = htons(0xffff);
 	struct ethtool_rx_flow_spec_input input = {};
-	__be16 vlan_tci = 0, vlan_m_tci = 0xffff;
 	unsigned int slice_num, rule_index[2];
 	const struct cfp_udf_layout *layout;
 	struct ethtool_rx_flow_rule *flow;
-- 
2.27.0.rc2

[PATCH net-next 3/3] dsa: bmc_sf2: Pass GENMASK() signed bits

[Andrew Lunn]

Oddly, GENMASK() requires signed bit numbers, so that it can compare
them for < 0. If passed an unsigned type, we get warnings about the
test never being true. There is no danger of overflow here, udf is
always a u8, so there is plenty of space when expanding to an int.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/bcm_sf2_cfp.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dsa/bcm_sf2_cfp.c b/drivers/net/dsa/bcm_sf2_cfp.c
index b54339477853..d82cee5d9202 100644
--- a/drivers/net/dsa/bcm_sf2_cfp.c
+++ b/drivers/net/dsa/bcm_sf2_cfp.c
@@ -128,12 +128,12 @@ static inline unsigned int bcm_sf2_get_num_udf_slices(const u8 *layout)
 	return count;
 }
 
-static inline u32 udf_upper_bits(unsigned int num_udf)
+static inline u32 udf_upper_bits(int num_udf)
 {
 	return GENMASK(num_udf - 1, 0) >> (UDFS_PER_SLICE - 1);
 }
 
-static inline u32 udf_lower_bits(unsigned int num_udf)
+static inline u32 udf_lower_bits(int num_udf)
 {
 	return (u8)GENMASK(num_udf - 1, 0);
 }
-- 
2.27.0.rc2

Re: [PATCH net-next 2/3] dsa: bcm_sf2: Initialize __be16 with a __be16 value

[Florian Fainelli]

On 7/5/2020 1:36 PM, Andrew Lunn wrote:
> A __be16 variable should be initialised with a __be16 value.  So add a
> htons(). In this case it is pointless, given the value being assigned
> is 0xffff, but it stops sparse from warnings.
> 
> Signed-off-by: Andrew Lunn <andrew@lunn.ch>

The subject should be:

net: dsa: bcm_sf2: Initialize __be16 with a __be16 value

to be consistent with other commits done to the same file, with that:

Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
-- 
Florian

Re: [PATCH net-next 1/3] net: dsa: b53: Fixup endianness warnings

[Florian Fainelli]

On 7/5/2020 1:36 PM, Andrew Lunn wrote:
> leX_to_cpu() expects to be passed an __leX type.
> 
> Signed-off-by: Andrew Lunn <andrew@lunn.ch>

Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
-- 
Florian

Re: [PATCH net-next 3/3] dsa: bmc_sf2: Pass GENMASK() signed bits

[Florian Fainelli]

On 7/5/2020 1:36 PM, Andrew Lunn wrote:
> Oddly, GENMASK() requires signed bit numbers, so that it can compare
> them for < 0. If passed an unsigned type, we get warnings about the
> test never being true. There is no danger of overflow here, udf is
> always a u8, so there is plenty of space when expanding to an int.
> 
> Signed-off-by: Andrew Lunn <andrew@lunn.ch>

Likewise for this patch, the subject should be:

net: dsa: bcm_sf2: Pass GENMASK() signed bits

(not the typo on bmc_sf2 vs. bcm_sf2), with that fixed:

Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
-- 
Florian

Re: [PATCH net-next 0/3] dsa: b53/sf2

[David Miller]

From: Andrew Lunn <andrew@lunn.ch>
Date: Sun,  5 Jul 2020 22:36:22 +0200

> Fixup most of the C=1 W=1 warnings in these drivers.

Series applied with patch #1 and #2's subjects fixed.