netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net-next v6 00/25] Multiple improvement to qca8k stability
@ 2021-05-14 20:59 Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant Ansuel Smith
                   ` (25 more replies)
  0 siblings, 26 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Currently qca8337 switch are widely used on ipq8064 based router.
On these particular router it was notice a very unstable switch with
port not link detected as link with unknown speed, port dropping
randomly and general unreliability. Lots of testing and comparison
between this dsa driver and the original qsdk driver showed lack of some
additional delay and values. A main difference arised from the original
driver and the dsa one. The original driver didn't use MASTER regs to
read phy status and the dedicated mdio driver worked correctly. Now that
the dsa driver actually use these regs, it was found that these special
read/write operation required mutual exclusion to normal
qca8k_read/write operation. The add of mutex for these operation fixed
the random port dropping and now only the actual linked port randomly
dropped. Adding additional delay for set_page operation and fixing a bug
in the mdio dedicated driver fixed also this problem. The current driver
requires also more time to apply vlan switch. All of these changes and
tweak permit a now very stable and reliable dsa driver and 0 port
dropping. This series is currently tested by at least 5 user with
different routers and all reports positive results and no problems.

Changes v6:
- Fix spelling mistake
- Change ms to ns (confirmed by datasheet)
Changes v5:
- Removed mdio patch (sent separetly to try to reduce the series)
  I know it was asked to reduced this series since it big, but rework
  the new changes to skip and error check looks wrong. Since half of them
  are actually already reviewed I think it's better to keep this series as is.
- Improve rgmii configurable patch
- Move qca8k phy dedicated driver to at803x phy driver
- Add support for dedicated internal mdio driver for qca8k
Changes v4:
- Use iopoll for busy_wait function
- Better describe and split some confusing commits
- Fix bad rgmii delay configurable patch
- Drop phy generic patch to pass flags with phylink_connect_phy
- Add dsa2 patch to declare mdio node in the switch node
- Add dsa patch to permit dsa driver to declare custom get_phys_mii_mask
    Some background about the last 2 patch.
    The qca8k switch doesn't have a 1:1 map between port reg and phy reg.
    Currently it's used a function to convert port to the internal phy reg.
    I added some patch to fix this.
    - The dsa driver now check if the mdio node is present and use the of variant
      of the mdiobus_register
    - A custom phy_mii_mask is required as currently the mask is generated from
      the port reg, but in our case the mask would be different as it should be
      generated from the phy reg. To generalize this I added an extra function
      that driver can provide to pass custom phy_mii_mask.
Changes v3:
- Revert mdio writel changes (use regmap with REGCACHE disabled)
- Split propagate error patch to 4 different patch
Changes v2:
- Implemented phy driver for internal PHYs
  I'm testing cable test functions as I found some documentation that
  actually declare regs about it. Problem is that it doesn't actually
  work. It seems that the value set are ignored by the phy.
- Made the rgmii delay configurable
- Reordered patch
- Split mdio patches to more specific ones
- Reworked mdio driver to use readl/writel instead of regmap
- Reworked the entire driver to make it aware of any read/write error.
- Added phy generic patch to pass flags with phylink_connect_phy
  function

Ansuel Smith (25):
  net: dsa: qca8k: change simple print to dev variant
  net: dsa: qca8k: use iopoll macro for qca8k_busy_wait
  net: dsa: qca8k: improve qca8k read/write/rmw bus access
  net: dsa: qca8k: handle qca8k_set_page errors
  net: dsa: qca8k: handle error with qca8k_read operation
  net: dsa: qca8k: handle error with qca8k_write operation
  net: dsa: qca8k: handle error with qca8k_rmw operation
  net: dsa: qca8k: handle error from qca8k_busy_wait
  net: dsa: qca8k: add support for qca8327 switch
  devicetree: net: dsa: qca8k: Document new compatible qca8327
  net: dsa: qca8k: add priority tweak to qca8337 switch
  net: dsa: qca8k: limit port5 delay to qca8337
  net: dsa: qca8k: add GLOBAL_FC settings needed for qca8327
  net: dsa: qca8k: add support for switch rev
  net: dsa: qca8k: add ethernet-ports fallback to setup_mdio_bus
  net: dsa: qca8k: make rgmii delay configurable
  net: dsa: qca8k: clear MASTER_EN after phy read/write
  net: dsa: qca8k: dsa: qca8k: protect MASTER busy_wait with mdio mutex
  net: dsa: qca8k: enlarge mdio delay and timeout
  net: dsa: qca8k: add support for internal phy and internal mdio
  devicetree: bindings: dsa: qca8k: Document internal mdio definition
  net: dsa: qca8k: improve internal mdio read/write bus access
  net: dsa: qca8k: pass switch_revision info to phy dev_flags
  net: phy: at803x: clean whitespace errors
  net: phy: add support for qca8k switch internal PHY in at803x

 .../devicetree/bindings/net/dsa/qca8k.txt     |  40 +
 drivers/net/dsa/qca8k.c                       | 758 ++++++++++++++----
 drivers/net/dsa/qca8k.h                       |  58 +-
 drivers/net/phy/Kconfig                       |   5 +-
 drivers/net/phy/at803x.c                      | 162 +++-
 5 files changed, 835 insertions(+), 188 deletions(-)

-- 
2.30.2


^ permalink raw reply	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 22:43   ` Russell King (Oracle)
  2021-05-14 20:59 ` [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait Ansuel Smith
                   ` (24 subsequent siblings)
  25 siblings, 1 reply; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Change pr_err and pr_warn to dev variant.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index cdaf9f85a2cb..0b295da6c356 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -701,7 +701,7 @@ qca8k_setup(struct dsa_switch *ds)
 
 	/* Make sure that port 0 is the cpu port */
 	if (!dsa_is_cpu_port(ds, 0)) {
-		pr_err("port 0 is not the CPU port\n");
+		dev_err(priv->dev, "port 0 is not the CPU port");
 		return -EINVAL;
 	}
 
@@ -711,7 +711,7 @@ qca8k_setup(struct dsa_switch *ds)
 	priv->regmap = devm_regmap_init(ds->dev, NULL, priv,
 					&qca8k_regmap_config);
 	if (IS_ERR(priv->regmap))
-		pr_warn("regmap initialization failed");
+		dev_warn(priv->dev, "regmap initialization failed");
 
 	ret = qca8k_setup_mdio_bus(priv);
 	if (ret)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 22:52   ` Russell King (Oracle)
  2021-05-14 20:59 ` [PATCH net-next v6 03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access Ansuel Smith
                   ` (23 subsequent siblings)
  25 siblings, 1 reply; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Use iopoll macro instead of while loop.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 23 +++++++++++------------
 drivers/net/dsa/qca8k.h |  2 ++
 2 files changed, 13 insertions(+), 12 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 0b295da6c356..25fa7084e820 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -262,21 +262,20 @@ static struct regmap_config qca8k_regmap_config = {
 static int
 qca8k_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
 {
-	unsigned long timeout;
-
-	timeout = jiffies + msecs_to_jiffies(20);
+	u32 val;
+	int ret;
 
-	/* loop until the busy flag has cleared */
-	do {
-		u32 val = qca8k_read(priv, reg);
-		int busy = val & mask;
+	ret = read_poll_timeout(qca8k_read, val, !(val & mask),
+				0, QCA8K_BUSY_WAIT_TIMEOUT * USEC_PER_MSEC, false,
+				priv, reg);
 
-		if (!busy)
-			break;
-		cond_resched();
-	} while (!time_after_eq(jiffies, timeout));
+	/* Check if qca8k_read has failed for a different reason
+	 * before returning -ETIMEDOUT
+	 */
+	if (ret < 0 && val < 0)
+		return val;
 
-	return time_after_eq(jiffies, timeout);
+	return ret;
 }
 
 static void
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index 7ca4b93e0bb5..86c585b7ec4a 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -18,6 +18,8 @@
 #define PHY_ID_QCA8337					0x004dd036
 #define QCA8K_ID_QCA8337				0x13
 
+#define QCA8K_BUSY_WAIT_TIMEOUT				20
+
 #define QCA8K_NUM_FDB_RECORDS				2048
 
 #define QCA8K_CPU_PORT					0
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 22:53   ` Russell King (Oracle)
  2021-05-14 20:59 ` [PATCH net-next v6 04/25] net: dsa: qca8k: handle qca8k_set_page errors Ansuel Smith
                   ` (22 subsequent siblings)
  25 siblings, 1 reply; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Put bus in local variable to improve faster access to the mdio bus.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 29 ++++++++++++++++-------------
 1 file changed, 16 insertions(+), 13 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 25fa7084e820..3c882d325fdf 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -142,17 +142,18 @@ qca8k_set_page(struct mii_bus *bus, u16 page)
 static u32
 qca8k_read(struct qca8k_priv *priv, u32 reg)
 {
+	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
 	u32 val;
 
 	qca8k_split_addr(reg, &r1, &r2, &page);
 
-	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	qca8k_set_page(priv->bus, page);
-	val = qca8k_mii_read32(priv->bus, 0x10 | r2, r1);
+	qca8k_set_page(bus, page);
+	val = qca8k_mii_read32(bus, 0x10 | r2, r1);
 
-	mutex_unlock(&priv->bus->mdio_lock);
+	mutex_unlock(&bus->mdio_lock);
 
 	return val;
 }
@@ -160,35 +161,37 @@ qca8k_read(struct qca8k_priv *priv, u32 reg)
 static void
 qca8k_write(struct qca8k_priv *priv, u32 reg, u32 val)
 {
+	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
 
 	qca8k_split_addr(reg, &r1, &r2, &page);
 
-	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	qca8k_set_page(priv->bus, page);
-	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
+	qca8k_set_page(bus, page);
+	qca8k_mii_write32(bus, 0x10 | r2, r1, val);
 
-	mutex_unlock(&priv->bus->mdio_lock);
+	mutex_unlock(&bus->mdio_lock);
 }
 
 static u32
 qca8k_rmw(struct qca8k_priv *priv, u32 reg, u32 mask, u32 val)
 {
+	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
 	u32 ret;
 
 	qca8k_split_addr(reg, &r1, &r2, &page);
 
-	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	qca8k_set_page(priv->bus, page);
-	ret = qca8k_mii_read32(priv->bus, 0x10 | r2, r1);
+	qca8k_set_page(bus, page);
+	ret = qca8k_mii_read32(bus, 0x10 | r2, r1);
 	ret &= ~mask;
 	ret |= val;
-	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, ret);
+	qca8k_mii_write32(bus, 0x10 | r2, r1, ret);
 
-	mutex_unlock(&priv->bus->mdio_lock);
+	mutex_unlock(&bus->mdio_lock);
 
 	return ret;
 }
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 04/25] net: dsa: qca8k: handle qca8k_set_page errors
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (2 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 22:55   ` Russell King (Oracle)
  2021-05-14 20:59 ` [PATCH net-next v6 05/25] net: dsa: qca8k: handle error with qca8k_read operation Ansuel Smith
                   ` (21 subsequent siblings)
  25 siblings, 1 reply; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

With a remote possibility, the set_page function can fail. Since this is
a critical part of the write/read qca8k regs, propagate the error and
terminate any read/write operation.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 33 ++++++++++++++++++++++++++-------
 1 file changed, 26 insertions(+), 7 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 3c882d325fdf..c9830286fd6d 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -127,16 +127,23 @@ qca8k_mii_write32(struct mii_bus *bus, int phy_id, u32 regnum, u32 val)
 				    "failed to write qca8k 32bit register\n");
 }
 
-static void
+static int
 qca8k_set_page(struct mii_bus *bus, u16 page)
 {
+	int ret;
+
 	if (page == qca8k_current_page)
-		return;
+		return 0;
 
-	if (bus->write(bus, 0x18, 0, page) < 0)
+	ret = bus->write(bus, 0x18, 0, page);
+	if (ret < 0) {
 		dev_err_ratelimited(&bus->dev,
 				    "failed to set qca8k page\n");
+		return ret;
+	}
+
 	qca8k_current_page = page;
+	return 0;
 }
 
 static u32
@@ -150,11 +157,14 @@ qca8k_read(struct qca8k_priv *priv, u32 reg)
 
 	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	qca8k_set_page(bus, page);
+	val = qca8k_set_page(bus, page);
+	if (val < 0)
+		goto exit;
+
 	val = qca8k_mii_read32(bus, 0x10 | r2, r1);
 
+exit:
 	mutex_unlock(&bus->mdio_lock);
-
 	return val;
 }
 
@@ -163,14 +173,19 @@ qca8k_write(struct qca8k_priv *priv, u32 reg, u32 val)
 {
 	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
+	int ret;
 
 	qca8k_split_addr(reg, &r1, &r2, &page);
 
 	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	qca8k_set_page(bus, page);
+	ret = qca8k_set_page(bus, page);
+	if (ret < 0)
+		goto exit;
+
 	qca8k_mii_write32(bus, 0x10 | r2, r1, val);
 
+exit:
 	mutex_unlock(&bus->mdio_lock);
 }
 
@@ -185,12 +200,16 @@ qca8k_rmw(struct qca8k_priv *priv, u32 reg, u32 mask, u32 val)
 
 	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	qca8k_set_page(bus, page);
+	ret = qca8k_set_page(bus, page);
+	if (ret < 0)
+		goto exit;
+
 	ret = qca8k_mii_read32(bus, 0x10 | r2, r1);
 	ret &= ~mask;
 	ret |= val;
 	qca8k_mii_write32(bus, 0x10 | r2, r1, ret);
 
+exit:
 	mutex_unlock(&bus->mdio_lock);
 
 	return ret;
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 05/25] net: dsa: qca8k: handle error with qca8k_read operation
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (3 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 04/25] net: dsa: qca8k: handle qca8k_set_page errors Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 23:01   ` Russell King (Oracle)
  2021-05-14 20:59 ` [PATCH net-next v6 06/25] net: dsa: qca8k: handle error with qca8k_write operation Ansuel Smith
                   ` (20 subsequent siblings)
  25 siblings, 1 reply; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

qca8k_read can fail. Rework any user to handle error values and
correctly return.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 73 ++++++++++++++++++++++++++++++++---------
 1 file changed, 58 insertions(+), 15 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index c9830286fd6d..5eb4d13fe0ba 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -231,8 +231,13 @@ static int
 qca8k_regmap_read(void *ctx, uint32_t reg, uint32_t *val)
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ctx;
+	int ret;
+
+	ret = qca8k_read(priv, reg);
+	if (ret < 0)
+		return ret;
 
-	*val = qca8k_read(priv, reg);
+	*val = ret;
 
 	return 0;
 }
@@ -300,15 +305,20 @@ qca8k_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
 	return ret;
 }
 
-static void
+static int
 qca8k_fdb_read(struct qca8k_priv *priv, struct qca8k_fdb *fdb)
 {
-	u32 reg[4];
+	u32 reg[4], val;
 	int i;
 
 	/* load the ARL table into an array */
-	for (i = 0; i < 4; i++)
-		reg[i] = qca8k_read(priv, QCA8K_REG_ATU_DATA0 + (i * 4));
+	for (i = 0; i < 4; i++) {
+		val = qca8k_read(priv, QCA8K_REG_ATU_DATA0 + (i * 4));
+		if (val < 0)
+			return val;
+
+		reg[i] = val;
+	}
 
 	/* vid - 83:72 */
 	fdb->vid = (reg[2] >> QCA8K_ATU_VID_S) & QCA8K_ATU_VID_M;
@@ -323,6 +333,8 @@ qca8k_fdb_read(struct qca8k_priv *priv, struct qca8k_fdb *fdb)
 	fdb->mac[3] = (reg[0] >> QCA8K_ATU_ADDR3_S) & 0xff;
 	fdb->mac[4] = (reg[0] >> QCA8K_ATU_ADDR4_S) & 0xff;
 	fdb->mac[5] = reg[0] & 0xff;
+
+	return 0;
 }
 
 static void
@@ -374,6 +386,8 @@ qca8k_fdb_access(struct qca8k_priv *priv, enum qca8k_fdb_cmd cmd, int port)
 	/* Check for table full violation when adding an entry */
 	if (cmd == QCA8K_FDB_LOAD) {
 		reg = qca8k_read(priv, QCA8K_REG_ATU_FUNC);
+		if (reg < 0)
+			return reg;
 		if (reg & QCA8K_ATU_FUNC_FULL)
 			return -1;
 	}
@@ -388,10 +402,10 @@ qca8k_fdb_next(struct qca8k_priv *priv, struct qca8k_fdb *fdb, int port)
 
 	qca8k_fdb_write(priv, fdb->vid, fdb->port_mask, fdb->mac, fdb->aging);
 	ret = qca8k_fdb_access(priv, QCA8K_FDB_NEXT, port);
-	if (ret >= 0)
-		qca8k_fdb_read(priv, fdb);
+	if (ret < 0)
+		return ret;
 
-	return ret;
+	return qca8k_fdb_read(priv, fdb);
 }
 
 static int
@@ -449,6 +463,8 @@ qca8k_vlan_access(struct qca8k_priv *priv, enum qca8k_vlan_cmd cmd, u16 vid)
 	/* Check for table full violation when adding an entry */
 	if (cmd == QCA8K_VLAN_LOAD) {
 		reg = qca8k_read(priv, QCA8K_REG_VTU_FUNC1);
+		if (reg < 0)
+			return reg;
 		if (reg & QCA8K_VTU_FUNC1_FULL)
 			return -ENOMEM;
 	}
@@ -475,6 +491,8 @@ qca8k_vlan_add(struct qca8k_priv *priv, u8 port, u16 vid, bool untagged)
 		goto out;
 
 	reg = qca8k_read(priv, QCA8K_REG_VTU_FUNC0);
+	if (reg < 0)
+		return reg;
 	reg |= QCA8K_VTU_FUNC0_VALID | QCA8K_VTU_FUNC0_IVL_EN;
 	reg &= ~(QCA8K_VTU_FUNC0_EG_MODE_MASK << QCA8K_VTU_FUNC0_EG_MODE_S(port));
 	if (untagged)
@@ -506,6 +524,8 @@ qca8k_vlan_del(struct qca8k_priv *priv, u8 port, u16 vid)
 		goto out;
 
 	reg = qca8k_read(priv, QCA8K_REG_VTU_FUNC0);
+	if (reg < 0)
+		return reg;
 	reg &= ~(3 << QCA8K_VTU_FUNC0_EG_MODE_S(port));
 	reg |= QCA8K_VTU_FUNC0_EG_MODE_NOT <<
 			QCA8K_VTU_FUNC0_EG_MODE_S(port);
@@ -621,8 +641,11 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 			    QCA8K_MDIO_MASTER_BUSY))
 		return -ETIMEDOUT;
 
-	val = (qca8k_read(priv, QCA8K_MDIO_MASTER_CTRL) &
-		QCA8K_MDIO_MASTER_DATA_MASK);
+	val = qca8k_read(priv, QCA8K_MDIO_MASTER_CTRL);
+	if (val < 0)
+		return val;
+
+	val &= QCA8K_MDIO_MASTER_DATA_MASK;
 
 	return val;
 }
@@ -978,6 +1001,8 @@ qca8k_phylink_mac_link_state(struct dsa_switch *ds, int port,
 	u32 reg;
 
 	reg = qca8k_read(priv, QCA8K_REG_PORT_STATUS(port));
+	if (reg < 0)
+		return reg;
 
 	state->link = !!(reg & QCA8K_PORT_STATUS_LINK_UP);
 	state->an_complete = state->link;
@@ -1078,18 +1103,26 @@ qca8k_get_ethtool_stats(struct dsa_switch *ds, int port,
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
 	const struct qca8k_mib_desc *mib;
-	u32 reg, i;
+	u32 reg, i, val;
 	u64 hi;
 
 	for (i = 0; i < ARRAY_SIZE(ar8327_mib); i++) {
 		mib = &ar8327_mib[i];
 		reg = QCA8K_PORT_MIB_COUNTER(port) + mib->offset;
 
-		data[i] = qca8k_read(priv, reg);
+		val = qca8k_read(priv, reg);
+		if (val < 0)
+			continue;
+
 		if (mib->size == 2) {
 			hi = qca8k_read(priv, reg + 4);
-			data[i] |= hi << 32;
+			if (hi < 0)
+				continue;
 		}
+
+		data[i] = val;
+		if (mib->size == 2)
+			data[i] |= hi << 32;
 	}
 }
 
@@ -1107,18 +1140,25 @@ qca8k_set_mac_eee(struct dsa_switch *ds, int port, struct ethtool_eee *eee)
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
 	u32 lpi_en = QCA8K_REG_EEE_CTRL_LPI_EN(port);
+	int ret = 0;
 	u32 reg;
 
 	mutex_lock(&priv->reg_mutex);
 	reg = qca8k_read(priv, QCA8K_REG_EEE_CTRL);
+	if (reg < 0) {
+		ret = reg;
+		goto exit;
+	}
+
 	if (eee->eee_enabled)
 		reg |= lpi_en;
 	else
 		reg &= ~lpi_en;
 	qca8k_write(priv, QCA8K_REG_EEE_CTRL, reg);
-	mutex_unlock(&priv->reg_mutex);
 
-	return 0;
+exit:
+	mutex_unlock(&priv->reg_mutex);
+	return ret;
 }
 
 static int
@@ -1443,6 +1483,9 @@ qca8k_sw_probe(struct mdio_device *mdiodev)
 
 	/* read the switches ID register */
 	id = qca8k_read(priv, QCA8K_REG_MASK_CTRL);
+	if (id < 0)
+		return id;
+
 	id >>= QCA8K_MASK_CTRL_ID_S;
 	id &= QCA8K_MASK_CTRL_ID_M;
 	if (id != QCA8K_ID_QCA8337)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 06/25] net: dsa: qca8k: handle error with qca8k_write operation
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (4 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 05/25] net: dsa: qca8k: handle error with qca8k_read operation Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 07/25] net: dsa: qca8k: handle error with qca8k_rmw operation Ansuel Smith
                   ` (19 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

qca8k_write can fail. Rework any user to handle error values and
correctly return.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 102 ++++++++++++++++++++++++++--------------
 1 file changed, 67 insertions(+), 35 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 5eb4d13fe0ba..2fdd7c2e74d5 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -168,7 +168,7 @@ qca8k_read(struct qca8k_priv *priv, u32 reg)
 	return val;
 }
 
-static void
+static int
 qca8k_write(struct qca8k_priv *priv, u32 reg, u32 val)
 {
 	struct mii_bus *bus = priv->bus;
@@ -187,6 +187,7 @@ qca8k_write(struct qca8k_priv *priv, u32 reg, u32 val)
 
 exit:
 	mutex_unlock(&bus->mdio_lock);
+	return ret;
 }
 
 static u32
@@ -247,9 +248,7 @@ qca8k_regmap_write(void *ctx, uint32_t reg, uint32_t val)
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ctx;
 
-	qca8k_write(priv, reg, val);
-
-	return 0;
+	return qca8k_write(priv, reg, val);
 }
 
 static const struct regmap_range qca8k_readable_ranges[] = {
@@ -367,6 +366,7 @@ static int
 qca8k_fdb_access(struct qca8k_priv *priv, enum qca8k_fdb_cmd cmd, int port)
 {
 	u32 reg;
+	int ret;
 
 	/* Set the command and FDB index */
 	reg = QCA8K_ATU_FUNC_BUSY;
@@ -377,7 +377,9 @@ qca8k_fdb_access(struct qca8k_priv *priv, enum qca8k_fdb_cmd cmd, int port)
 	}
 
 	/* Write the function register triggering the table access */
-	qca8k_write(priv, QCA8K_REG_ATU_FUNC, reg);
+	ret = qca8k_write(priv, QCA8K_REG_ATU_FUNC, reg);
+	if (ret)
+		return ret;
 
 	/* wait for completion */
 	if (qca8k_busy_wait(priv, QCA8K_REG_ATU_FUNC, QCA8K_ATU_FUNC_BUSY))
@@ -447,6 +449,7 @@ static int
 qca8k_vlan_access(struct qca8k_priv *priv, enum qca8k_vlan_cmd cmd, u16 vid)
 {
 	u32 reg;
+	int ret;
 
 	/* Set the command and VLAN index */
 	reg = QCA8K_VTU_FUNC1_BUSY;
@@ -454,7 +457,9 @@ qca8k_vlan_access(struct qca8k_priv *priv, enum qca8k_vlan_cmd cmd, u16 vid)
 	reg |= vid << QCA8K_VTU_FUNC1_VID_S;
 
 	/* Write the function register triggering the table access */
-	qca8k_write(priv, QCA8K_REG_VTU_FUNC1, reg);
+	ret = qca8k_write(priv, QCA8K_REG_VTU_FUNC1, reg);
+	if (ret)
+		return ret;
 
 	/* wait for completion */
 	if (qca8k_busy_wait(priv, QCA8K_REG_VTU_FUNC1, QCA8K_VTU_FUNC1_BUSY))
@@ -502,7 +507,9 @@ qca8k_vlan_add(struct qca8k_priv *priv, u8 port, u16 vid, bool untagged)
 		reg |= QCA8K_VTU_FUNC0_EG_MODE_TAG <<
 				QCA8K_VTU_FUNC0_EG_MODE_S(port);
 
-	qca8k_write(priv, QCA8K_REG_VTU_FUNC0, reg);
+	ret = qca8k_write(priv, QCA8K_REG_VTU_FUNC0, reg);
+	if (ret)
+		return ret;
 	ret = qca8k_vlan_access(priv, QCA8K_VLAN_LOAD, vid);
 
 out:
@@ -545,7 +552,9 @@ qca8k_vlan_del(struct qca8k_priv *priv, u8 port, u16 vid)
 	if (del) {
 		ret = qca8k_vlan_access(priv, QCA8K_VLAN_PURGE, vid);
 	} else {
-		qca8k_write(priv, QCA8K_REG_VTU_FUNC0, reg);
+		ret = qca8k_write(priv, QCA8K_REG_VTU_FUNC0, reg);
+		if (ret)
+			return ret;
 		ret = qca8k_vlan_access(priv, QCA8K_VLAN_LOAD, vid);
 	}
 
@@ -555,15 +564,20 @@ qca8k_vlan_del(struct qca8k_priv *priv, u8 port, u16 vid)
 	return ret;
 }
 
-static void
+static int
 qca8k_mib_init(struct qca8k_priv *priv)
 {
+	int ret;
+
 	mutex_lock(&priv->reg_mutex);
 	qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_FLUSH | QCA8K_MIB_BUSY);
 	qca8k_busy_wait(priv, QCA8K_REG_MIB, QCA8K_MIB_BUSY);
 	qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_CPU_KEEP);
-	qca8k_write(priv, QCA8K_REG_MODULE_EN, QCA8K_MODULE_EN_MIB);
+
+	ret = qca8k_write(priv, QCA8K_REG_MODULE_EN, QCA8K_MODULE_EN_MIB);
+
 	mutex_unlock(&priv->reg_mutex);
+	return ret;
 }
 
 static void
@@ -600,6 +614,7 @@ static int
 qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 {
 	u32 phy, val;
+	int ret;
 
 	if (regnum >= QCA8K_MDIO_MASTER_MAX_REG)
 		return -EINVAL;
@@ -613,7 +628,9 @@ qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 	      QCA8K_MDIO_MASTER_REG_ADDR(regnum) |
 	      QCA8K_MDIO_MASTER_DATA(data);
 
-	qca8k_write(priv, QCA8K_MDIO_MASTER_CTRL, val);
+	ret = qca8k_write(priv, QCA8K_MDIO_MASTER_CTRL, val);
+	if (ret)
+		return ret;
 
 	return qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
 		QCA8K_MDIO_MASTER_BUSY);
@@ -623,6 +640,7 @@ static int
 qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 {
 	u32 phy, val;
+	int ret;
 
 	if (regnum >= QCA8K_MDIO_MASTER_MAX_REG)
 		return -EINVAL;
@@ -635,7 +653,9 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 	      QCA8K_MDIO_MASTER_READ | QCA8K_MDIO_MASTER_PHY_ADDR(phy) |
 	      QCA8K_MDIO_MASTER_REG_ADDR(regnum);
 
-	qca8k_write(priv, QCA8K_MDIO_MASTER_CTRL, val);
+	ret = qca8k_write(priv, QCA8K_MDIO_MASTER_CTRL, val);
+	if (ret)
+		return ret;
 
 	if (qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
 			    QCA8K_MDIO_MASTER_BUSY))
@@ -766,12 +786,18 @@ qca8k_setup(struct dsa_switch *ds)
 		      QCA8K_GLOBAL_FW_CTRL0_CPU_PORT_EN);
 
 	/* Enable MIB counters */
-	qca8k_mib_init(priv);
+	ret = qca8k_mib_init(priv);
+	if (ret)
+		dev_warn(priv->dev, "mib init failed");
 
 	/* Enable QCA header mode on the cpu port */
-	qca8k_write(priv, QCA8K_REG_PORT_HDR_CTRL(QCA8K_CPU_PORT),
-		    QCA8K_PORT_HDR_CTRL_ALL << QCA8K_PORT_HDR_CTRL_TX_S |
-		    QCA8K_PORT_HDR_CTRL_ALL << QCA8K_PORT_HDR_CTRL_RX_S);
+	ret = qca8k_write(priv, QCA8K_REG_PORT_HDR_CTRL(QCA8K_CPU_PORT),
+			  QCA8K_PORT_HDR_CTRL_ALL << QCA8K_PORT_HDR_CTRL_TX_S |
+			  QCA8K_PORT_HDR_CTRL_ALL << QCA8K_PORT_HDR_CTRL_RX_S);
+	if (ret) {
+		dev_err(priv->dev, "failed enabling QCA header mode");
+		return ret;
+	}
 
 	/* Disable forwarding by default on all ports */
 	for (i = 0; i < QCA8K_NUM_PORTS; i++)
@@ -783,11 +809,13 @@ qca8k_setup(struct dsa_switch *ds)
 		qca8k_port_set_status(priv, i, 0);
 
 	/* Forward all unknown frames to CPU port for Linux processing */
-	qca8k_write(priv, QCA8K_REG_GLOBAL_FW_CTRL1,
-		    BIT(0) << QCA8K_GLOBAL_FW_CTRL1_IGMP_DP_S |
-		    BIT(0) << QCA8K_GLOBAL_FW_CTRL1_BC_DP_S |
-		    BIT(0) << QCA8K_GLOBAL_FW_CTRL1_MC_DP_S |
-		    BIT(0) << QCA8K_GLOBAL_FW_CTRL1_UC_DP_S);
+	ret = qca8k_write(priv, QCA8K_REG_GLOBAL_FW_CTRL1,
+			  BIT(0) << QCA8K_GLOBAL_FW_CTRL1_IGMP_DP_S |
+			  BIT(0) << QCA8K_GLOBAL_FW_CTRL1_BC_DP_S |
+			  BIT(0) << QCA8K_GLOBAL_FW_CTRL1_MC_DP_S |
+			  BIT(0) << QCA8K_GLOBAL_FW_CTRL1_UC_DP_S);
+	if (ret)
+		return ret;
 
 	/* Setup connection between CPU port & user ports */
 	for (i = 0; i < QCA8K_NUM_PORTS; i++) {
@@ -815,16 +843,20 @@ qca8k_setup(struct dsa_switch *ds)
 			qca8k_rmw(priv, QCA8K_EGRESS_VLAN(i),
 				  0xfff << shift,
 				  QCA8K_PORT_VID_DEF << shift);
-			qca8k_write(priv, QCA8K_REG_PORT_VLAN_CTRL0(i),
-				    QCA8K_PORT_VLAN_CVID(QCA8K_PORT_VID_DEF) |
-				    QCA8K_PORT_VLAN_SVID(QCA8K_PORT_VID_DEF));
+			ret = qca8k_write(priv, QCA8K_REG_PORT_VLAN_CTRL0(i),
+					  QCA8K_PORT_VLAN_CVID(QCA8K_PORT_VID_DEF) |
+					  QCA8K_PORT_VLAN_SVID(QCA8K_PORT_VID_DEF));
+			if (ret)
+				return ret;
 		}
 	}
 
 	/* Setup our port MTUs to match power on defaults */
 	for (i = 0; i < QCA8K_NUM_PORTS; i++)
 		priv->port_mtu[i] = ETH_FRAME_LEN + ETH_FCS_LEN;
-	qca8k_write(priv, QCA8K_MAX_FRAME_SIZE, ETH_FRAME_LEN + ETH_FCS_LEN);
+	ret = qca8k_write(priv, QCA8K_MAX_FRAME_SIZE, ETH_FRAME_LEN + ETH_FCS_LEN);
+	if (ret)
+		dev_warn(priv->dev, "failed setting MTU settings");
 
 	/* Flush the FDB table */
 	qca8k_fdb_flush(priv);
@@ -1140,8 +1172,8 @@ qca8k_set_mac_eee(struct dsa_switch *ds, int port, struct ethtool_eee *eee)
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
 	u32 lpi_en = QCA8K_REG_EEE_CTRL_LPI_EN(port);
-	int ret = 0;
 	u32 reg;
+	int ret;
 
 	mutex_lock(&priv->reg_mutex);
 	reg = qca8k_read(priv, QCA8K_REG_EEE_CTRL);
@@ -1154,7 +1186,7 @@ qca8k_set_mac_eee(struct dsa_switch *ds, int port, struct ethtool_eee *eee)
 		reg |= lpi_en;
 	else
 		reg &= ~lpi_en;
-	qca8k_write(priv, QCA8K_REG_EEE_CTRL, reg);
+	ret = qca8k_write(priv, QCA8K_REG_EEE_CTRL, reg);
 
 exit:
 	mutex_unlock(&priv->reg_mutex);
@@ -1284,9 +1316,7 @@ qca8k_port_change_mtu(struct dsa_switch *ds, int port, int new_mtu)
 			mtu = priv->port_mtu[i];
 
 	/* Include L2 header / FCS length */
-	qca8k_write(priv, QCA8K_MAX_FRAME_SIZE, mtu + ETH_HLEN + ETH_FCS_LEN);
-
-	return 0;
+	return qca8k_write(priv, QCA8K_MAX_FRAME_SIZE, mtu + ETH_HLEN + ETH_FCS_LEN);
 }
 
 static int
@@ -1381,7 +1411,7 @@ qca8k_port_vlan_add(struct dsa_switch *ds, int port,
 	bool untagged = vlan->flags & BRIDGE_VLAN_INFO_UNTAGGED;
 	bool pvid = vlan->flags & BRIDGE_VLAN_INFO_PVID;
 	struct qca8k_priv *priv = ds->priv;
-	int ret = 0;
+	int ret;
 
 	ret = qca8k_vlan_add(priv, port, vlan->vid, untagged);
 	if (ret) {
@@ -1394,9 +1424,11 @@ qca8k_port_vlan_add(struct dsa_switch *ds, int port,
 
 		qca8k_rmw(priv, QCA8K_EGRESS_VLAN(port),
 			  0xfff << shift, vlan->vid << shift);
-		qca8k_write(priv, QCA8K_REG_PORT_VLAN_CTRL0(port),
-			    QCA8K_PORT_VLAN_CVID(vlan->vid) |
-			    QCA8K_PORT_VLAN_SVID(vlan->vid));
+		ret = qca8k_write(priv, QCA8K_REG_PORT_VLAN_CTRL0(port),
+				  QCA8K_PORT_VLAN_CVID(vlan->vid) |
+				  QCA8K_PORT_VLAN_SVID(vlan->vid));
+		if (ret)
+			return ret;
 	}
 
 	return 0;
@@ -1407,7 +1439,7 @@ qca8k_port_vlan_del(struct dsa_switch *ds, int port,
 		    const struct switchdev_obj_port_vlan *vlan)
 {
 	struct qca8k_priv *priv = ds->priv;
-	int ret = 0;
+	int ret;
 
 	ret = qca8k_vlan_del(priv, port, vlan->vid);
 	if (ret)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 07/25] net: dsa: qca8k: handle error with qca8k_rmw operation
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (5 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 06/25] net: dsa: qca8k: handle error with qca8k_write operation Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 08/25] net: dsa: qca8k: handle error from qca8k_busy_wait Ansuel Smith
                   ` (18 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

qca8k_rmw can fail. Rework any user to handle error values and
correctly return. Change qca8k_rmw to return the error code or 0 instead
of the reg value. The reg returned by qca8k_rmw wasn't used anywhere,
so this doesn't cause any functional change.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 133 +++++++++++++++++++++++++---------------
 1 file changed, 83 insertions(+), 50 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 2fdd7c2e74d5..409f6592048a 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -190,12 +190,13 @@ qca8k_write(struct qca8k_priv *priv, u32 reg, u32 val)
 	return ret;
 }
 
-static u32
-qca8k_rmw(struct qca8k_priv *priv, u32 reg, u32 mask, u32 val)
+static int
+qca8k_rmw(struct qca8k_priv *priv, u32 reg, u32 mask, u32 write_val)
 {
 	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
-	u32 ret;
+	u32 val;
+	int ret;
 
 	qca8k_split_addr(reg, &r1, &r2, &page);
 
@@ -205,10 +206,15 @@ qca8k_rmw(struct qca8k_priv *priv, u32 reg, u32 mask, u32 val)
 	if (ret < 0)
 		goto exit;
 
-	ret = qca8k_mii_read32(bus, 0x10 | r2, r1);
-	ret &= ~mask;
-	ret |= val;
-	qca8k_mii_write32(bus, 0x10 | r2, r1, ret);
+	val = qca8k_mii_read32(bus, 0x10 | r2, r1);
+	if (val < 0) {
+		ret = val;
+		goto exit;
+	}
+
+	val &= ~mask;
+	val |= write_val;
+	qca8k_mii_write32(bus, 0x10 | r2, r1, val);
 
 exit:
 	mutex_unlock(&bus->mdio_lock);
@@ -216,16 +222,16 @@ qca8k_rmw(struct qca8k_priv *priv, u32 reg, u32 mask, u32 val)
 	return ret;
 }
 
-static void
+static int
 qca8k_reg_set(struct qca8k_priv *priv, u32 reg, u32 val)
 {
-	qca8k_rmw(priv, reg, 0, val);
+	return qca8k_rmw(priv, reg, 0, val);
 }
 
-static void
+static int
 qca8k_reg_clear(struct qca8k_priv *priv, u32 reg, u32 val)
 {
-	qca8k_rmw(priv, reg, val, 0);
+	return qca8k_rmw(priv, reg, val, 0);
 }
 
 static int
@@ -570,12 +576,19 @@ qca8k_mib_init(struct qca8k_priv *priv)
 	int ret;
 
 	mutex_lock(&priv->reg_mutex);
-	qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_FLUSH | QCA8K_MIB_BUSY);
+	ret = qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_FLUSH | QCA8K_MIB_BUSY);
+	if (ret)
+		goto exit;
+
 	qca8k_busy_wait(priv, QCA8K_REG_MIB, QCA8K_MIB_BUSY);
-	qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_CPU_KEEP);
+
+	ret = qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_CPU_KEEP);
+	if (ret)
+		goto exit;
 
 	ret = qca8k_write(priv, QCA8K_REG_MODULE_EN, QCA8K_MODULE_EN_MIB);
 
+exit:
 	mutex_unlock(&priv->reg_mutex);
 	return ret;
 }
@@ -747,9 +760,8 @@ qca8k_setup_mdio_bus(struct qca8k_priv *priv)
 		 * a dt-overlay and driver reload changed the configuration
 		 */
 
-		qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
-				QCA8K_MDIO_MASTER_EN);
-		return 0;
+		return qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
+				       QCA8K_MDIO_MASTER_EN);
 	}
 
 	priv->ops.phy_read = qca8k_phy_read;
@@ -782,8 +794,12 @@ qca8k_setup(struct dsa_switch *ds)
 		return ret;
 
 	/* Enable CPU Port */
-	qca8k_reg_set(priv, QCA8K_REG_GLOBAL_FW_CTRL0,
-		      QCA8K_GLOBAL_FW_CTRL0_CPU_PORT_EN);
+	ret = qca8k_reg_set(priv, QCA8K_REG_GLOBAL_FW_CTRL0,
+			    QCA8K_GLOBAL_FW_CTRL0_CPU_PORT_EN);
+	if (ret) {
+		dev_err(priv->dev, "failed enabling CPU port");
+		return ret;
+	}
 
 	/* Enable MIB counters */
 	ret = qca8k_mib_init(priv);
@@ -800,9 +816,12 @@ qca8k_setup(struct dsa_switch *ds)
 	}
 
 	/* Disable forwarding by default on all ports */
-	for (i = 0; i < QCA8K_NUM_PORTS; i++)
-		qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(i),
-			  QCA8K_PORT_LOOKUP_MEMBER, 0);
+	for (i = 0; i < QCA8K_NUM_PORTS; i++) {
+		ret = qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(i),
+				QCA8K_PORT_LOOKUP_MEMBER, 0);
+		if (ret)
+			return ret;
+	}
 
 	/* Disable MAC by default on all ports */
 	for (i = 1; i < QCA8K_NUM_PORTS; i++)
@@ -821,28 +840,37 @@ qca8k_setup(struct dsa_switch *ds)
 	for (i = 0; i < QCA8K_NUM_PORTS; i++) {
 		/* CPU port gets connected to all user ports of the switch */
 		if (dsa_is_cpu_port(ds, i)) {
-			qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(QCA8K_CPU_PORT),
-				  QCA8K_PORT_LOOKUP_MEMBER, dsa_user_ports(ds));
+			ret = qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(QCA8K_CPU_PORT),
+					QCA8K_PORT_LOOKUP_MEMBER, dsa_user_ports(ds));
+			if (ret)
+				return ret;
 		}
 
 		/* Individual user ports get connected to CPU port only */
 		if (dsa_is_user_port(ds, i)) {
 			int shift = 16 * (i % 2);
 
-			qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(i),
-				  QCA8K_PORT_LOOKUP_MEMBER,
-				  BIT(QCA8K_CPU_PORT));
+			ret = qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(i),
+					QCA8K_PORT_LOOKUP_MEMBER,
+					BIT(QCA8K_CPU_PORT));
+			if (ret)
+				return ret;
 
 			/* Enable ARP Auto-learning by default */
-			qca8k_reg_set(priv, QCA8K_PORT_LOOKUP_CTRL(i),
-				      QCA8K_PORT_LOOKUP_LEARN);
+			ret = qca8k_reg_set(priv, QCA8K_PORT_LOOKUP_CTRL(i),
+					    QCA8K_PORT_LOOKUP_LEARN);
+			if (ret)
+				return ret;
 
 			/* For port based vlans to work we need to set the
 			 * default egress vid
 			 */
-			qca8k_rmw(priv, QCA8K_EGRESS_VLAN(i),
-				  0xfff << shift,
-				  QCA8K_PORT_VID_DEF << shift);
+			ret = qca8k_rmw(priv, QCA8K_EGRESS_VLAN(i),
+					0xfff << shift,
+					QCA8K_PORT_VID_DEF << shift);
+			if (ret)
+				return ret;
+
 			ret = qca8k_write(priv, QCA8K_REG_PORT_VLAN_CTRL0(i),
 					  QCA8K_PORT_VLAN_CVID(QCA8K_PORT_VID_DEF) |
 					  QCA8K_PORT_VLAN_SVID(QCA8K_PORT_VID_DEF));
@@ -1234,7 +1262,7 @@ qca8k_port_bridge_join(struct dsa_switch *ds, int port, struct net_device *br)
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
 	int port_mask = BIT(QCA8K_CPU_PORT);
-	int i;
+	int i, ret;
 
 	for (i = 1; i < QCA8K_NUM_PORTS; i++) {
 		if (dsa_to_port(ds, i)->bridge_dev != br)
@@ -1242,17 +1270,20 @@ qca8k_port_bridge_join(struct dsa_switch *ds, int port, struct net_device *br)
 		/* Add this port to the portvlan mask of the other ports
 		 * in the bridge
 		 */
-		qca8k_reg_set(priv,
-			      QCA8K_PORT_LOOKUP_CTRL(i),
-			      BIT(port));
+		ret = qca8k_reg_set(priv,
+				    QCA8K_PORT_LOOKUP_CTRL(i),
+				    BIT(port));
+		if (ret)
+			return ret;
 		if (i != port)
 			port_mask |= BIT(i);
 	}
+
 	/* Add all other ports to this ports portvlan mask */
-	qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(port),
-		  QCA8K_PORT_LOOKUP_MEMBER, port_mask);
+	ret = qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(port),
+			QCA8K_PORT_LOOKUP_MEMBER, port_mask);
 
-	return 0;
+	return ret;
 }
 
 static void
@@ -1389,18 +1420,19 @@ qca8k_port_vlan_filtering(struct dsa_switch *ds, int port, bool vlan_filtering,
 			  struct netlink_ext_ack *extack)
 {
 	struct qca8k_priv *priv = ds->priv;
+	int ret;
 
 	if (vlan_filtering) {
-		qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(port),
-			  QCA8K_PORT_LOOKUP_VLAN_MODE,
-			  QCA8K_PORT_LOOKUP_VLAN_MODE_SECURE);
+		ret = qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(port),
+				QCA8K_PORT_LOOKUP_VLAN_MODE,
+				QCA8K_PORT_LOOKUP_VLAN_MODE_SECURE);
 	} else {
-		qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(port),
-			  QCA8K_PORT_LOOKUP_VLAN_MODE,
-			  QCA8K_PORT_LOOKUP_VLAN_MODE_NONE);
+		ret = qca8k_rmw(priv, QCA8K_PORT_LOOKUP_CTRL(port),
+				QCA8K_PORT_LOOKUP_VLAN_MODE,
+				QCA8K_PORT_LOOKUP_VLAN_MODE_NONE);
 	}
 
-	return 0;
+	return ret;
 }
 
 static int
@@ -1422,16 +1454,17 @@ qca8k_port_vlan_add(struct dsa_switch *ds, int port,
 	if (pvid) {
 		int shift = 16 * (port % 2);
 
-		qca8k_rmw(priv, QCA8K_EGRESS_VLAN(port),
-			  0xfff << shift, vlan->vid << shift);
+		ret = qca8k_rmw(priv, QCA8K_EGRESS_VLAN(port),
+				0xfff << shift, vlan->vid << shift);
+		if (ret)
+			return ret;
+
 		ret = qca8k_write(priv, QCA8K_REG_PORT_VLAN_CTRL0(port),
 				  QCA8K_PORT_VLAN_CVID(vlan->vid) |
 				  QCA8K_PORT_VLAN_SVID(vlan->vid));
-		if (ret)
-			return ret;
 	}
 
-	return 0;
+	return ret;
 }
 
 static int
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 08/25] net: dsa: qca8k: handle error from qca8k_busy_wait
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (6 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 07/25] net: dsa: qca8k: handle error with qca8k_rmw operation Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 20:59 ` [PATCH net-next v6 09/25] net: dsa: qca8k: add support for qca8327 switch Ansuel Smith
                   ` (17 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Propagate errors from qca8k_busy_wait instead of hardcoding return
value.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 409f6592048a..d4e3f81576ec 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -388,8 +388,9 @@ qca8k_fdb_access(struct qca8k_priv *priv, enum qca8k_fdb_cmd cmd, int port)
 		return ret;
 
 	/* wait for completion */
-	if (qca8k_busy_wait(priv, QCA8K_REG_ATU_FUNC, QCA8K_ATU_FUNC_BUSY))
-		return -1;
+	ret = qca8k_busy_wait(priv, QCA8K_REG_ATU_FUNC, QCA8K_ATU_FUNC_BUSY);
+	if (ret)
+		return ret;
 
 	/* Check for table full violation when adding an entry */
 	if (cmd == QCA8K_FDB_LOAD) {
@@ -468,8 +469,9 @@ qca8k_vlan_access(struct qca8k_priv *priv, enum qca8k_vlan_cmd cmd, u16 vid)
 		return ret;
 
 	/* wait for completion */
-	if (qca8k_busy_wait(priv, QCA8K_REG_VTU_FUNC1, QCA8K_VTU_FUNC1_BUSY))
-		return -ETIMEDOUT;
+	ret = qca8k_busy_wait(priv, QCA8K_REG_VTU_FUNC1, QCA8K_VTU_FUNC1_BUSY);
+	if (ret)
+		return ret;
 
 	/* Check for table full violation when adding an entry */
 	if (cmd == QCA8K_VLAN_LOAD) {
@@ -580,7 +582,9 @@ qca8k_mib_init(struct qca8k_priv *priv)
 	if (ret)
 		goto exit;
 
-	qca8k_busy_wait(priv, QCA8K_REG_MIB, QCA8K_MIB_BUSY);
+	ret = qca8k_busy_wait(priv, QCA8K_REG_MIB, QCA8K_MIB_BUSY);
+	if (ret)
+		goto exit;
 
 	ret = qca8k_reg_set(priv, QCA8K_REG_MIB, QCA8K_MIB_CPU_KEEP);
 	if (ret)
@@ -670,9 +674,10 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 	if (ret)
 		return ret;
 
-	if (qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
-			    QCA8K_MDIO_MASTER_BUSY))
-		return -ETIMEDOUT;
+	ret = qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
+			      QCA8K_MDIO_MASTER_BUSY);
+	if (ret)
+		return ret;
 
 	val = qca8k_read(priv, QCA8K_MDIO_MASTER_CTRL);
 	if (val < 0)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 09/25] net: dsa: qca8k: add support for qca8327 switch
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (7 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 08/25] net: dsa: qca8k: handle error from qca8k_busy_wait Ansuel Smith
@ 2021-05-14 20:59 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 10/25] devicetree: net: dsa: qca8k: Document new compatible qca8327 Ansuel Smith
                   ` (16 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 20:59 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

qca8327 switch is a low tier version of the more recent qca8337.
It does share the same regs used by the qca8k driver and can be
supported with minimal change.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/net/dsa/qca8k.c | 23 ++++++++++++++++++++---
 drivers/net/dsa/qca8k.h |  6 ++++++
 2 files changed, 26 insertions(+), 3 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index d4e3f81576ec..693bd9fd532b 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -1524,6 +1524,7 @@ static const struct dsa_switch_ops qca8k_switch_ops = {
 static int
 qca8k_sw_probe(struct mdio_device *mdiodev)
 {
+	const struct qca8k_match_data *data;
 	struct qca8k_priv *priv;
 	u32 id;
 
@@ -1551,6 +1552,11 @@ qca8k_sw_probe(struct mdio_device *mdiodev)
 		gpiod_set_value_cansleep(priv->reset_gpio, 0);
 	}
 
+	/* get the switches ID from the compatible */
+	data = of_device_get_match_data(&mdiodev->dev);
+	if (!data)
+		return -ENODEV;
+
 	/* read the switches ID register */
 	id = qca8k_read(priv, QCA8K_REG_MASK_CTRL);
 	if (id < 0)
@@ -1558,8 +1564,10 @@ qca8k_sw_probe(struct mdio_device *mdiodev)
 
 	id >>= QCA8K_MASK_CTRL_ID_S;
 	id &= QCA8K_MASK_CTRL_ID_M;
-	if (id != QCA8K_ID_QCA8337)
+	if (id != data->id) {
+		dev_err(&mdiodev->dev, "Switch id detected %x but expected %x", id, data->id);
 		return -ENODEV;
+	}
 
 	priv->ds = devm_kzalloc(&mdiodev->dev, sizeof(*priv->ds), GFP_KERNEL);
 	if (!priv->ds)
@@ -1624,9 +1632,18 @@ static int qca8k_resume(struct device *dev)
 static SIMPLE_DEV_PM_OPS(qca8k_pm_ops,
 			 qca8k_suspend, qca8k_resume);
 
+static const struct qca8k_match_data qca832x = {
+	.id = QCA8K_ID_QCA8327,
+};
+
+static const struct qca8k_match_data qca833x = {
+	.id = QCA8K_ID_QCA8337,
+};
+
 static const struct of_device_id qca8k_of_match[] = {
-	{ .compatible = "qca,qca8334" },
-	{ .compatible = "qca,qca8337" },
+	{ .compatible = "qca,qca8327", .data = &qca832x },
+	{ .compatible = "qca,qca8334", .data = &qca833x },
+	{ .compatible = "qca,qca8337", .data = &qca833x },
 	{ /* sentinel */ },
 };
 
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index 86c585b7ec4a..87a8b10459c6 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -15,6 +15,8 @@
 #define QCA8K_NUM_PORTS					7
 #define QCA8K_MAX_MTU					9000
 
+#define PHY_ID_QCA8327					0x004dd034
+#define QCA8K_ID_QCA8327				0x12
 #define PHY_ID_QCA8337					0x004dd036
 #define QCA8K_ID_QCA8337				0x13
 
@@ -213,6 +215,10 @@ struct ar8xxx_port_status {
 	int enabled;
 };
 
+struct qca8k_match_data {
+	u8 id;
+};
+
 struct qca8k_priv {
 	struct regmap *regmap;
 	struct mii_bus *bus;
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 10/25] devicetree: net: dsa: qca8k: Document new compatible qca8327
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (8 preceding siblings ...)
  2021-05-14 20:59 ` [PATCH net-next v6 09/25] net: dsa: qca8k: add support for qca8327 switch Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 11/25] net: dsa: qca8k: add priority tweak to qca8337 switch Ansuel Smith
                   ` (15 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith, Rob Herring

Add support for qca8327 in the compatible list.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
---
 Documentation/devicetree/bindings/net/dsa/qca8k.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/net/dsa/qca8k.txt b/Documentation/devicetree/bindings/net/dsa/qca8k.txt
index ccbc6d89325d..1daf68e7ae19 100644
--- a/Documentation/devicetree/bindings/net/dsa/qca8k.txt
+++ b/Documentation/devicetree/bindings/net/dsa/qca8k.txt
@@ -3,6 +3,7 @@
 Required properties:
 
 - compatible: should be one of:
+    "qca,qca8327"
     "qca,qca8334"
     "qca,qca8337"
 
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 11/25] net: dsa: qca8k: add priority tweak to qca8337 switch
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (9 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 10/25] devicetree: net: dsa: qca8k: Document new compatible qca8327 Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 12/25] net: dsa: qca8k: limit port5 delay to qca8337 Ansuel Smith
                   ` (14 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

The port 5 of the qca8337 have some problem in flood condition. The
original legacy driver had some specific buffer and priority settings
for the different port suggested by the QCA switch team. Add this
missing settings to improve switch stability under load condition.
The packet priority tweak is only needed for the qca8337 switch and
other qca8k switch are not affected.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 47 +++++++++++++++++++++++++++++++++++++++++
 drivers/net/dsa/qca8k.h | 25 ++++++++++++++++++++++
 2 files changed, 72 insertions(+)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 693bd9fd532b..65f27d136aef 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -779,6 +779,7 @@ qca8k_setup(struct dsa_switch *ds)
 {
 	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
 	int ret, i;
+	u32 mask;
 
 	/* Make sure that port 0 is the cpu port */
 	if (!dsa_is_cpu_port(ds, 0)) {
@@ -884,6 +885,51 @@ qca8k_setup(struct dsa_switch *ds)
 		}
 	}
 
+	/* The port 5 of the qca8337 have some problem in flood condition. The
+	 * original legacy driver had some specific buffer and priority settings
+	 * for the different port suggested by the QCA switch team. Add this
+	 * missing settings to improve switch stability under load condition.
+	 * This problem is limited to qca8337 and other qca8k switch are not affected.
+	 */
+	if (priv->switch_id == QCA8K_ID_QCA8337) {
+		for (i = 0; i < QCA8K_NUM_PORTS; i++) {
+			switch (i) {
+			/* The 2 CPU port and port 5 requires some different
+			 * priority than any other ports.
+			 */
+			case 0:
+			case 5:
+			case 6:
+				mask = QCA8K_PORT_HOL_CTRL0_EG_PRI0(0x3) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI1(0x4) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI2(0x4) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI3(0x4) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI4(0x6) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI5(0x8) |
+					QCA8K_PORT_HOL_CTRL0_EG_PORT(0x1e);
+				break;
+			default:
+				mask = QCA8K_PORT_HOL_CTRL0_EG_PRI0(0x3) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI1(0x4) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI2(0x6) |
+					QCA8K_PORT_HOL_CTRL0_EG_PRI3(0x8) |
+					QCA8K_PORT_HOL_CTRL0_EG_PORT(0x19);
+			}
+			qca8k_write(priv, QCA8K_REG_PORT_HOL_CTRL0(i), mask);
+
+			mask = QCA8K_PORT_HOL_CTRL1_ING(0x6) |
+			QCA8K_PORT_HOL_CTRL1_EG_PRI_BUF_EN |
+			QCA8K_PORT_HOL_CTRL1_EG_PORT_BUF_EN |
+			QCA8K_PORT_HOL_CTRL1_WRED_EN;
+			qca8k_rmw(priv, QCA8K_REG_PORT_HOL_CTRL1(i),
+				  QCA8K_PORT_HOL_CTRL1_ING_BUF |
+				  QCA8K_PORT_HOL_CTRL1_EG_PRI_BUF_EN |
+				  QCA8K_PORT_HOL_CTRL1_EG_PORT_BUF_EN |
+				  QCA8K_PORT_HOL_CTRL1_WRED_EN,
+				  mask);
+		}
+	}
+
 	/* Setup our port MTUs to match power on defaults */
 	for (i = 0; i < QCA8K_NUM_PORTS; i++)
 		priv->port_mtu[i] = ETH_FRAME_LEN + ETH_FCS_LEN;
@@ -1569,6 +1615,7 @@ qca8k_sw_probe(struct mdio_device *mdiodev)
 		return -ENODEV;
 	}
 
+	priv->switch_id = id;
 	priv->ds = devm_kzalloc(&mdiodev->dev, sizeof(*priv->ds), GFP_KERNEL);
 	if (!priv->ds)
 		return -ENOMEM;
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index 87a8b10459c6..42d90836dffa 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -168,6 +168,30 @@
 #define   QCA8K_PORT_LOOKUP_STATE			GENMASK(18, 16)
 #define   QCA8K_PORT_LOOKUP_LEARN			BIT(20)
 
+#define QCA8K_REG_PORT_HOL_CTRL0(_i)			(0x970 + (_i) * 0x8)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI0_BUF		GENMASK(3, 0)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI0(x)		((x) << 0)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI1_BUF		GENMASK(7, 4)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI1(x)		((x) << 4)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI2_BUF		GENMASK(11, 8)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI2(x)		((x) << 8)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI3_BUF		GENMASK(15, 12)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI3(x)		((x) << 12)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI4_BUF		GENMASK(19, 16)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI4(x)		((x) << 16)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI5_BUF		GENMASK(23, 20)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PRI5(x)		((x) << 20)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PORT_BUF		GENMASK(29, 24)
+#define   QCA8K_PORT_HOL_CTRL0_EG_PORT(x)		((x) << 24)
+
+#define QCA8K_REG_PORT_HOL_CTRL1(_i)			(0x974 + (_i) * 0x8)
+#define   QCA8K_PORT_HOL_CTRL1_ING_BUF			GENMASK(3, 0)
+#define   QCA8K_PORT_HOL_CTRL1_ING(x)			((x) << 0)
+#define   QCA8K_PORT_HOL_CTRL1_EG_PRI_BUF_EN		BIT(6)
+#define   QCA8K_PORT_HOL_CTRL1_EG_PORT_BUF_EN		BIT(7)
+#define   QCA8K_PORT_HOL_CTRL1_WRED_EN			BIT(8)
+#define   QCA8K_PORT_HOL_CTRL1_EG_MIRROR_EN		BIT(16)
+
 /* Pkt edit registers */
 #define QCA8K_EGRESS_VLAN(x)				(0x0c70 + (4 * (x / 2)))
 
@@ -220,6 +244,7 @@ struct qca8k_match_data {
 };
 
 struct qca8k_priv {
+	u8 switch_id;
 	struct regmap *regmap;
 	struct mii_bus *bus;
 	struct ar8xxx_port_status port_sts[QCA8K_NUM_PORTS];
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 12/25] net: dsa: qca8k: limit port5 delay to qca8337
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (10 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 11/25] net: dsa: qca8k: add priority tweak to qca8337 switch Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 13/25] net: dsa: qca8k: add GLOBAL_FC settings needed for qca8327 Ansuel Smith
                   ` (13 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Limit port5 rx delay to qca8337. This is taken from the legacy QSDK code
that limits the rx delay on port5 to only this particular switch version,
on other switch only the tx and rx delay for port0 are needed.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 65f27d136aef..b598930190e1 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -1003,8 +1003,10 @@ qca8k_phylink_mac_config(struct dsa_switch *ds, int port, unsigned int mode,
 			    QCA8K_PORT_PAD_RGMII_EN |
 			    QCA8K_PORT_PAD_RGMII_TX_DELAY(QCA8K_MAX_DELAY) |
 			    QCA8K_PORT_PAD_RGMII_RX_DELAY(QCA8K_MAX_DELAY));
-		qca8k_write(priv, QCA8K_REG_PORT5_PAD_CTRL,
-			    QCA8K_PORT_PAD_RGMII_RX_DELAY_EN);
+		/* QCA8337 requires to set rgmii rx delay */
+		if (priv->switch_id == QCA8K_ID_QCA8337)
+			qca8k_write(priv, QCA8K_REG_PORT5_PAD_CTRL,
+				    QCA8K_PORT_PAD_RGMII_RX_DELAY_EN);
 		break;
 	case PHY_INTERFACE_MODE_SGMII:
 	case PHY_INTERFACE_MODE_1000BASEX:
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 13/25] net: dsa: qca8k: add GLOBAL_FC settings needed for qca8327
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (11 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 12/25] net: dsa: qca8k: limit port5 delay to qca8337 Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 14/25] net: dsa: qca8k: add support for switch rev Ansuel Smith
                   ` (12 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Switch qca8327 needs special settings for the GLOBAL_FC_THRES regs.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 10 ++++++++++
 drivers/net/dsa/qca8k.h |  6 ++++++
 2 files changed, 16 insertions(+)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index b598930190e1..10e3e1ca7e95 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -930,6 +930,16 @@ qca8k_setup(struct dsa_switch *ds)
 		}
 	}
 
+	/* Special GLOBAL_FC_THRESH value are needed for ar8327 switch */
+	if (priv->switch_id == QCA8K_ID_QCA8327) {
+		mask = QCA8K_GLOBAL_FC_GOL_XON_THRES(288) |
+		       QCA8K_GLOBAL_FC_GOL_XOFF_THRES(496);
+		qca8k_rmw(priv, QCA8K_REG_GLOBAL_FC_THRESH,
+			  QCA8K_GLOBAL_FC_GOL_XON_THRES_S |
+			  QCA8K_GLOBAL_FC_GOL_XOFF_THRES_S,
+			  mask);
+	}
+
 	/* Setup our port MTUs to match power on defaults */
 	for (i = 0; i < QCA8K_NUM_PORTS; i++)
 		priv->port_mtu[i] = ETH_FRAME_LEN + ETH_FCS_LEN;
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index 42d90836dffa..eceeacfe2c5d 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -168,6 +168,12 @@
 #define   QCA8K_PORT_LOOKUP_STATE			GENMASK(18, 16)
 #define   QCA8K_PORT_LOOKUP_LEARN			BIT(20)
 
+#define QCA8K_REG_GLOBAL_FC_THRESH			0x800
+#define   QCA8K_GLOBAL_FC_GOL_XON_THRES(x)		((x) << 16)
+#define   QCA8K_GLOBAL_FC_GOL_XON_THRES_S		GENMASK(24, 16)
+#define   QCA8K_GLOBAL_FC_GOL_XOFF_THRES(x)		((x) << 0)
+#define   QCA8K_GLOBAL_FC_GOL_XOFF_THRES_S		GENMASK(8, 0)
+
 #define QCA8K_REG_PORT_HOL_CTRL0(_i)			(0x970 + (_i) * 0x8)
 #define   QCA8K_PORT_HOL_CTRL0_EG_PRI0_BUF		GENMASK(3, 0)
 #define   QCA8K_PORT_HOL_CTRL0_EG_PRI0(x)		((x) << 0)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 14/25] net: dsa: qca8k: add support for switch rev
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (12 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 13/25] net: dsa: qca8k: add GLOBAL_FC settings needed for qca8327 Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 15/25] net: dsa: qca8k: add ethernet-ports fallback to setup_mdio_bus Ansuel Smith
                   ` (11 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

qca8k internal phy driver require some special debug value to be set
based on the switch revision. Rework the switch id read function to
also read the chip revision.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/net/dsa/qca8k.c | 53 ++++++++++++++++++++++++++---------------
 drivers/net/dsa/qca8k.h |  7 ++++--
 2 files changed, 39 insertions(+), 21 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 10e3e1ca7e95..35ff4cf08786 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -1579,12 +1579,40 @@ static const struct dsa_switch_ops qca8k_switch_ops = {
 	.phylink_mac_link_up	= qca8k_phylink_mac_link_up,
 };
 
+static int qca8k_read_switch_id(struct qca8k_priv *priv)
+{
+	const struct qca8k_match_data *data;
+	u32 val;
+	u8 id;
+
+	/* get the switches ID from the compatible */
+	data = of_device_get_match_data(priv->dev);
+	if (!data)
+		return -ENODEV;
+
+	val = qca8k_read(priv, QCA8K_REG_MASK_CTRL);
+	if (val < 0)
+		return -ENODEV;
+
+	id = QCA8K_MASK_CTRL_DEVICE_ID(val & QCA8K_MASK_CTRL_DEVICE_ID_MASK);
+	if (id != data->id) {
+		dev_err(priv->dev, "Switch id detected %x but expected %x", id, data->id);
+		return -ENODEV;
+	}
+
+	priv->switch_id = id;
+
+	/* Save revision to communicate to the internal PHY driver */
+	priv->switch_revision = (val & QCA8K_MASK_CTRL_REV_ID_MASK);
+
+	return 0;
+}
+
 static int
 qca8k_sw_probe(struct mdio_device *mdiodev)
 {
-	const struct qca8k_match_data *data;
 	struct qca8k_priv *priv;
-	u32 id;
+	int ret;
 
 	/* allocate the private data struct so that we can probe the switches
 	 * ID register
@@ -1610,24 +1638,11 @@ qca8k_sw_probe(struct mdio_device *mdiodev)
 		gpiod_set_value_cansleep(priv->reset_gpio, 0);
 	}
 
-	/* get the switches ID from the compatible */
-	data = of_device_get_match_data(&mdiodev->dev);
-	if (!data)
-		return -ENODEV;
-
-	/* read the switches ID register */
-	id = qca8k_read(priv, QCA8K_REG_MASK_CTRL);
-	if (id < 0)
-		return id;
-
-	id >>= QCA8K_MASK_CTRL_ID_S;
-	id &= QCA8K_MASK_CTRL_ID_M;
-	if (id != data->id) {
-		dev_err(&mdiodev->dev, "Switch id detected %x but expected %x", id, data->id);
-		return -ENODEV;
-	}
+	/* Check the detected switch id */
+	ret = qca8k_read_switch_id(priv);
+	if (ret)
+		return ret;
 
-	priv->switch_id = id;
 	priv->ds = devm_kzalloc(&mdiodev->dev, sizeof(*priv->ds), GFP_KERNEL);
 	if (!priv->ds)
 		return -ENOMEM;
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index eceeacfe2c5d..338277978ec0 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -30,8 +30,10 @@
 
 /* Global control registers */
 #define QCA8K_REG_MASK_CTRL				0x000
-#define   QCA8K_MASK_CTRL_ID_M				0xff
-#define   QCA8K_MASK_CTRL_ID_S				8
+#define   QCA8K_MASK_CTRL_REV_ID_MASK			GENMASK(7, 0)
+#define   QCA8K_MASK_CTRL_REV_ID(x)			((x) >> 0)
+#define   QCA8K_MASK_CTRL_DEVICE_ID_MASK		GENMASK(15, 8)
+#define   QCA8K_MASK_CTRL_DEVICE_ID(x)			((x) >> 8)
 #define QCA8K_REG_PORT0_PAD_CTRL			0x004
 #define QCA8K_REG_PORT5_PAD_CTRL			0x008
 #define QCA8K_REG_PORT6_PAD_CTRL			0x00c
@@ -251,6 +253,7 @@ struct qca8k_match_data {
 
 struct qca8k_priv {
 	u8 switch_id;
+	u8 switch_revision;
 	struct regmap *regmap;
 	struct mii_bus *bus;
 	struct ar8xxx_port_status port_sts[QCA8K_NUM_PORTS];
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 15/25] net: dsa: qca8k: add ethernet-ports fallback to setup_mdio_bus
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (13 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 14/25] net: dsa: qca8k: add support for switch rev Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 16/25] net: dsa: qca8k: make rgmii delay configurable Ansuel Smith
                   ` (10 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Dsa now also supports ethernet-ports. Add this new binding as a fallback
if the ports node can't be found.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 35ff4cf08786..cc9ab35f8b17 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -718,6 +718,9 @@ qca8k_setup_mdio_bus(struct qca8k_priv *priv)
 	int err;
 
 	ports = of_get_child_by_name(priv->dev->of_node, "ports");
+	if (!ports)
+		ports = of_get_child_by_name(priv->dev->of_node, "ethernet-ports");
+
 	if (!ports)
 		return -EINVAL;
 
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 16/25] net: dsa: qca8k: make rgmii delay configurable
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (14 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 15/25] net: dsa: qca8k: add ethernet-ports fallback to setup_mdio_bus Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 17/25] net: dsa: qca8k: clear MASTER_EN after phy read/write Ansuel Smith
                   ` (9 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

The legacy qsdk code used a different delay instead of the max value.
Qsdk use 1 ns for rx and 2 ns for tx. Make these values configurable
using the standard rx/tx-internal-delay-ps ethernet binding and apply
qsdk values by default. The connected gmac doesn't add any delay so no
additional delay is added to tx/rx.
On this switch the delay is actually in ns so value should be in the
1000 order. Any value converted from ps to ns by dividing it by 1000
as the switch max value for delay is 3ns.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 82 ++++++++++++++++++++++++++++++++++++++++-
 drivers/net/dsa/qca8k.h | 11 +++---
 2 files changed, 86 insertions(+), 7 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index cc9ab35f8b17..dedbc6565516 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -777,6 +777,68 @@ qca8k_setup_mdio_bus(struct qca8k_priv *priv)
 	return 0;
 }
 
+static int
+qca8k_setup_of_rgmii_delay(struct qca8k_priv *priv)
+{
+	struct device_node *port_dn;
+	phy_interface_t mode;
+	struct dsa_port *dp;
+	u32 val;
+
+	/* CPU port is already checked */
+	dp = dsa_to_port(priv->ds, 0);
+
+	port_dn = dp->dn;
+
+	/* Check if port 0 is set to the correct type */
+	of_get_phy_mode(port_dn, &mode);
+	if (mode != PHY_INTERFACE_MODE_RGMII_ID &&
+	    mode != PHY_INTERFACE_MODE_RGMII_RXID &&
+	    mode != PHY_INTERFACE_MODE_RGMII_TXID) {
+		return 0;
+	}
+
+	switch (mode) {
+	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
+		if (of_property_read_u32(port_dn, "rx-internal-delay-ps", &val))
+			val = 2;
+		else
+			/* Switch regs accept value in ns, convert ps to ns */
+			val = val / 1000;
+
+		if (val > QCA8K_MAX_DELAY) {
+			dev_err(priv->dev, "rgmii rx delay is limited to a max value of 3ns, setting to the max value");
+			val = 3;
+		}
+
+		priv->rgmii_rx_delay = val;
+		/* Stop here if we need to check only for rx delay */
+		if (mode != PHY_INTERFACE_MODE_RGMII_ID)
+			break;
+
+		fallthrough;
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		if (of_property_read_u32(port_dn, "tx-internal-delay-ps", &val))
+			val = 1;
+		else
+			/* Switch regs accept value in ns, convert ps to ns */
+			val = val / 1000;
+
+		if (val > QCA8K_MAX_DELAY) {
+			dev_err(priv->dev, "rgmii tx delay is limited to a max value of 3ns, setting to the max value");
+			val = 3;
+		}
+
+		priv->rgmii_tx_delay = val;
+		break;
+	default:
+		return 0;
+	}
+
+	return 0;
+}
+
 static int
 qca8k_setup(struct dsa_switch *ds)
 {
@@ -802,6 +864,10 @@ qca8k_setup(struct dsa_switch *ds)
 	if (ret)
 		return ret;
 
+	ret = qca8k_setup_of_rgmii_delay(priv);
+	if (ret)
+		return ret;
+
 	/* Enable CPU Port */
 	ret = qca8k_reg_set(priv, QCA8K_REG_GLOBAL_FW_CTRL0,
 			    QCA8K_GLOBAL_FW_CTRL0_CPU_PORT_EN);
@@ -970,6 +1036,8 @@ qca8k_phylink_mac_config(struct dsa_switch *ds, int port, unsigned int mode,
 	case 0: /* 1st CPU port */
 		if (state->interface != PHY_INTERFACE_MODE_RGMII &&
 		    state->interface != PHY_INTERFACE_MODE_RGMII_ID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_TXID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_RXID &&
 		    state->interface != PHY_INTERFACE_MODE_SGMII)
 			return;
 
@@ -985,6 +1053,8 @@ qca8k_phylink_mac_config(struct dsa_switch *ds, int port, unsigned int mode,
 	case 6: /* 2nd CPU port / external PHY */
 		if (state->interface != PHY_INTERFACE_MODE_RGMII &&
 		    state->interface != PHY_INTERFACE_MODE_RGMII_ID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_TXID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_RXID &&
 		    state->interface != PHY_INTERFACE_MODE_SGMII &&
 		    state->interface != PHY_INTERFACE_MODE_1000BASEX)
 			return;
@@ -1008,14 +1078,18 @@ qca8k_phylink_mac_config(struct dsa_switch *ds, int port, unsigned int mode,
 		qca8k_write(priv, reg, QCA8K_PORT_PAD_RGMII_EN);
 		break;
 	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
 		/* RGMII_ID needs internal delay. This is enabled through
 		 * PORT5_PAD_CTRL for all ports, rather than individual port
 		 * registers
 		 */
 		qca8k_write(priv, reg,
 			    QCA8K_PORT_PAD_RGMII_EN |
-			    QCA8K_PORT_PAD_RGMII_TX_DELAY(QCA8K_MAX_DELAY) |
-			    QCA8K_PORT_PAD_RGMII_RX_DELAY(QCA8K_MAX_DELAY));
+			    QCA8K_PORT_PAD_RGMII_TX_DELAY(priv->rgmii_tx_delay) |
+			    QCA8K_PORT_PAD_RGMII_RX_DELAY(priv->rgmii_rx_delay) |
+			    QCA8K_PORT_PAD_RGMII_TX_DELAY_EN |
+			    QCA8K_PORT_PAD_RGMII_RX_DELAY_EN);
 		/* QCA8337 requires to set rgmii rx delay */
 		if (priv->switch_id == QCA8K_ID_QCA8337)
 			qca8k_write(priv, QCA8K_REG_PORT5_PAD_CTRL,
@@ -1073,6 +1147,8 @@ qca8k_phylink_validate(struct dsa_switch *ds, int port,
 		if (state->interface != PHY_INTERFACE_MODE_NA &&
 		    state->interface != PHY_INTERFACE_MODE_RGMII &&
 		    state->interface != PHY_INTERFACE_MODE_RGMII_ID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_TXID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_RXID &&
 		    state->interface != PHY_INTERFACE_MODE_SGMII)
 			goto unsupported;
 		break;
@@ -1090,6 +1166,8 @@ qca8k_phylink_validate(struct dsa_switch *ds, int port,
 		if (state->interface != PHY_INTERFACE_MODE_NA &&
 		    state->interface != PHY_INTERFACE_MODE_RGMII &&
 		    state->interface != PHY_INTERFACE_MODE_RGMII_ID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_TXID &&
+		    state->interface != PHY_INTERFACE_MODE_RGMII_RXID &&
 		    state->interface != PHY_INTERFACE_MODE_SGMII &&
 		    state->interface != PHY_INTERFACE_MODE_1000BASEX)
 			goto unsupported;
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index 338277978ec0..a878486d9bcd 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -38,12 +38,11 @@
 #define QCA8K_REG_PORT5_PAD_CTRL			0x008
 #define QCA8K_REG_PORT6_PAD_CTRL			0x00c
 #define   QCA8K_PORT_PAD_RGMII_EN			BIT(26)
-#define   QCA8K_PORT_PAD_RGMII_TX_DELAY(x)		\
-						((0x8 + (x & 0x3)) << 22)
-#define   QCA8K_PORT_PAD_RGMII_RX_DELAY(x)		\
-						((0x10 + (x & 0x3)) << 20)
-#define   QCA8K_MAX_DELAY				3
+#define   QCA8K_PORT_PAD_RGMII_TX_DELAY(x)		((x) << 22)
+#define   QCA8K_PORT_PAD_RGMII_RX_DELAY(x)		((x) << 20)
+#define	  QCA8K_PORT_PAD_RGMII_TX_DELAY_EN		BIT(25)
 #define   QCA8K_PORT_PAD_RGMII_RX_DELAY_EN		BIT(24)
+#define   QCA8K_MAX_DELAY				3
 #define   QCA8K_PORT_PAD_SGMII_EN			BIT(7)
 #define QCA8K_REG_PWS					0x010
 #define   QCA8K_PWS_SERDES_AEN_DIS			BIT(7)
@@ -254,6 +253,8 @@ struct qca8k_match_data {
 struct qca8k_priv {
 	u8 switch_id;
 	u8 switch_revision;
+	u8 rgmii_tx_delay;
+	u8 rgmii_rx_delay;
 	struct regmap *regmap;
 	struct mii_bus *bus;
 	struct ar8xxx_port_status port_sts[QCA8K_NUM_PORTS];
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 17/25] net: dsa: qca8k: clear MASTER_EN after phy read/write
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (15 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 16/25] net: dsa: qca8k: make rgmii delay configurable Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 18/25] net: dsa: qca8k: dsa: qca8k: protect MASTER busy_wait with mdio mutex Ansuel Smith
                   ` (8 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Clear MDIO_MASTER_EN bit from MDIO_MASTER_CTRL after read/write
operation. The MDIO_MASTER_EN bit is not reset after read/write
operation and the next operation can be wrongly interpreted by the
switch as a mdio operation. This cause a production of wrong/garbage
data from the switch and underfined bheavior. (random port drop,
unplugged port flagged with link up, wrong port speed)
Also on driver remove the MASTER_CTRL can be left set and cause the
malfunction of any next driver using the mdio device.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index dedbc6565516..a2b4d5097868 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -649,8 +649,14 @@ qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 	if (ret)
 		return ret;
 
-	return qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
-		QCA8K_MDIO_MASTER_BUSY);
+	ret = qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
+			      QCA8K_MDIO_MASTER_BUSY);
+
+	/* even if the busy_wait timeouts try to clear the MASTER_EN */
+	qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
+			QCA8K_MDIO_MASTER_EN);
+
+	return ret;
 }
 
 static int
@@ -685,6 +691,10 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 
 	val &= QCA8K_MDIO_MASTER_DATA_MASK;
 
+	/* even if the busy_wait timeouts try to clear the MASTER_EN */
+	qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
+			QCA8K_MDIO_MASTER_EN);
+
 	return val;
 }
 
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 18/25] net: dsa: qca8k: dsa: qca8k: protect MASTER busy_wait with mdio mutex
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (16 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 17/25] net: dsa: qca8k: clear MASTER_EN after phy read/write Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 19/25] net: dsa: qca8k: enlarge mdio delay and timeout Ansuel Smith
                   ` (7 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

MDIO_MASTER operation have a dedicated busy wait that is not protected
by the mdio mutex. This can cause situation where the MASTER operation
is done and a normal operation is executed between the MASTER read/write
and the MASTER busy_wait. Rework the qca8k_mdio_read/write function to
address this issue by binding the lock for the whole MASTER operation
and not only the mdio read/write common operation.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 drivers/net/dsa/qca8k.c | 68 +++++++++++++++++++++++++++++++++--------
 1 file changed, 55 insertions(+), 13 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index a2b4d5097868..1f8bfe0a78f4 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -627,9 +627,32 @@ qca8k_port_to_phy(int port)
 	return port - 1;
 }
 
+static int
+qca8k_mdio_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
+{
+	u16 r1, r2, page;
+	u32 val;
+	int ret;
+
+	qca8k_split_addr(reg, &r1, &r2, &page);
+
+	ret = read_poll_timeout(qca8k_mii_read32, val, !(val & mask), 0,
+				QCA8K_BUSY_WAIT_TIMEOUT * USEC_PER_MSEC, false,
+				priv->bus, 0x10 | r2, r1);
+
+	/* Check if qca8k_read has failed for a different reason
+	 * before returnting -ETIMEDOUT
+	 */
+	if (ret < 0 && val < 0)
+		return val;
+
+	return ret;
+}
+
 static int
 qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 {
+	u16 r1, r2, page;
 	u32 phy, val;
 	int ret;
 
@@ -645,12 +668,21 @@ qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 	      QCA8K_MDIO_MASTER_REG_ADDR(regnum) |
 	      QCA8K_MDIO_MASTER_DATA(data);
 
-	ret = qca8k_write(priv, QCA8K_MDIO_MASTER_CTRL, val);
+	qca8k_split_addr(QCA8K_MDIO_MASTER_CTRL, &r1, &r2, &page);
+
+	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+
+	ret = qca8k_set_page(priv->bus, page);
 	if (ret)
-		return ret;
+		goto exit;
+
+	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
 
-	ret = qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
-			      QCA8K_MDIO_MASTER_BUSY);
+	ret = qca8k_mdio_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
+				   QCA8K_MDIO_MASTER_BUSY);
+
+exit:
+	mutex_unlock(&priv->bus->mdio_lock);
 
 	/* even if the busy_wait timeouts try to clear the MASTER_EN */
 	qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
@@ -662,6 +694,7 @@ qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 static int
 qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 {
+	u16 r1, r2, page;
 	u32 phy, val;
 	int ret;
 
@@ -676,21 +709,30 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 	      QCA8K_MDIO_MASTER_READ | QCA8K_MDIO_MASTER_PHY_ADDR(phy) |
 	      QCA8K_MDIO_MASTER_REG_ADDR(regnum);
 
-	ret = qca8k_write(priv, QCA8K_MDIO_MASTER_CTRL, val);
-	if (ret)
-		return ret;
+	qca8k_split_addr(QCA8K_MDIO_MASTER_CTRL, &r1, &r2, &page);
 
-	ret = qca8k_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
-			      QCA8K_MDIO_MASTER_BUSY);
+	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+
+	ret = qca8k_set_page(priv->bus, page);
 	if (ret)
-		return ret;
+		goto exit;
 
-	val = qca8k_read(priv, QCA8K_MDIO_MASTER_CTRL);
-	if (val < 0)
-		return val;
+	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
 
+	ret = qca8k_mdio_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
+				   QCA8K_MDIO_MASTER_BUSY);
+	if (ret)
+		goto exit;
+
+	val = qca8k_mii_read32(priv->bus, 0x10 | r2, r1);
 	val &= QCA8K_MDIO_MASTER_DATA_MASK;
 
+exit:
+	mutex_unlock(&priv->bus->mdio_lock);
+
+	if (val >= 0)
+		val &= QCA8K_MDIO_MASTER_DATA_MASK;
+
 	/* even if the busy_wait timeouts try to clear the MASTER_EN */
 	qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
 			QCA8K_MDIO_MASTER_EN);
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 19/25] net: dsa: qca8k: enlarge mdio delay and timeout
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (17 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 18/25] net: dsa: qca8k: dsa: qca8k: protect MASTER busy_wait with mdio mutex Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 20/25] net: dsa: qca8k: add support for internal phy and internal mdio Ansuel Smith
                   ` (6 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

The witch require some extra delay after setting page or the next
read/write can use still use the old page. Add a delay after the
set_page function to address this as it's done in QSDK legacy driver.
Some timeouts were notice with VLAN and phy function, enlarge the
mdio busy wait timeout to fix these problems.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 1 +
 drivers/net/dsa/qca8k.h | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index 1f8bfe0a78f4..df4cf6d75074 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -143,6 +143,7 @@ qca8k_set_page(struct mii_bus *bus, u16 page)
 	}
 
 	qca8k_current_page = page;
+	usleep_range(1000, 2000);
 	return 0;
 }
 
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index a878486d9bcd..d365f85ab34f 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -20,7 +20,7 @@
 #define PHY_ID_QCA8337					0x004dd036
 #define QCA8K_ID_QCA8337				0x13
 
-#define QCA8K_BUSY_WAIT_TIMEOUT				20
+#define QCA8K_BUSY_WAIT_TIMEOUT				2000
 
 #define QCA8K_NUM_FDB_RECORDS				2048
 
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 20/25] net: dsa: qca8k: add support for internal phy and internal mdio
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (18 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 19/25] net: dsa: qca8k: enlarge mdio delay and timeout Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 21/25] devicetree: bindings: dsa: qca8k: Document internal mdio definition Ansuel Smith
                   ` (5 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Add support to setup_mdio_bus for internal phy declaration. Introduce a
flag to use the legacy port phy mapping by default and use the direct
mapping if a mdio node is detected in the switch node. Register a
dedicated mdio internal mdio bus to address the different mapping
between port and phy if the mdio node is detected.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 112 +++++++++++++++++++++++++++++-----------
 drivers/net/dsa/qca8k.h |   1 +
 2 files changed, 83 insertions(+), 30 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index df4cf6d75074..ba288181fd1a 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -11,6 +11,7 @@
 #include <linux/netdevice.h>
 #include <net/dsa.h>
 #include <linux/of_net.h>
+#include <linux/of_mdio.h>
 #include <linux/of_platform.h>
 #include <linux/if_bridge.h>
 #include <linux/mdio.h>
@@ -629,7 +630,7 @@ qca8k_port_to_phy(int port)
 }
 
 static int
-qca8k_mdio_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
+qca8k_mdio_busy_wait(struct mii_bus *bus, u32 reg, u32 mask)
 {
 	u16 r1, r2, page;
 	u32 val;
@@ -639,7 +640,7 @@ qca8k_mdio_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
 
 	ret = read_poll_timeout(qca8k_mii_read32, val, !(val & mask), 0,
 				QCA8K_BUSY_WAIT_TIMEOUT * USEC_PER_MSEC, false,
-				priv->bus, 0x10 | r2, r1);
+				bus, 0x10 | r2, r1);
 
 	/* Check if qca8k_read has failed for a different reason
 	 * before returnting -ETIMEDOUT
@@ -651,19 +652,16 @@ qca8k_mdio_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
 }
 
 static int
-qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
+qca8k_mdio_write(struct mii_bus *salve_bus, int phy, int regnum, u16 data)
 {
+	struct qca8k_priv *priv = salve_bus->priv;
 	u16 r1, r2, page;
-	u32 phy, val;
+	u32 val;
 	int ret;
 
 	if (regnum >= QCA8K_MDIO_MASTER_MAX_REG)
 		return -EINVAL;
 
-	/* callee is responsible for not passing bad ports,
-	 * but we still would like to make spills impossible.
-	 */
-	phy = qca8k_port_to_phy(port) % PHY_MAX_ADDR;
 	val = QCA8K_MDIO_MASTER_BUSY | QCA8K_MDIO_MASTER_EN |
 	      QCA8K_MDIO_MASTER_WRITE | QCA8K_MDIO_MASTER_PHY_ADDR(phy) |
 	      QCA8K_MDIO_MASTER_REG_ADDR(regnum) |
@@ -679,33 +677,29 @@ qca8k_mdio_write(struct qca8k_priv *priv, int port, u32 regnum, u16 data)
 
 	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
 
-	ret = qca8k_mdio_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
+	ret = qca8k_mdio_busy_wait(priv->bus, QCA8K_MDIO_MASTER_CTRL,
 				   QCA8K_MDIO_MASTER_BUSY);
 
 exit:
-	mutex_unlock(&priv->bus->mdio_lock);
-
 	/* even if the busy_wait timeouts try to clear the MASTER_EN */
-	qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
-			QCA8K_MDIO_MASTER_EN);
+	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, 0);
+
+	mutex_unlock(&priv->bus->mdio_lock);
 
 	return ret;
 }
 
 static int
-qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
+qca8k_mdio_read(struct mii_bus *salve_bus, int phy, int regnum)
 {
+	struct qca8k_priv *priv = salve_bus->priv;
 	u16 r1, r2, page;
-	u32 phy, val;
+	u32 val;
 	int ret;
 
 	if (regnum >= QCA8K_MDIO_MASTER_MAX_REG)
 		return -EINVAL;
 
-	/* callee is responsible for not passing bad ports,
-	 * but we still would like to make spills impossible.
-	 */
-	phy = qca8k_port_to_phy(port) % PHY_MAX_ADDR;
 	val = QCA8K_MDIO_MASTER_BUSY | QCA8K_MDIO_MASTER_EN |
 	      QCA8K_MDIO_MASTER_READ | QCA8K_MDIO_MASTER_PHY_ADDR(phy) |
 	      QCA8K_MDIO_MASTER_REG_ADDR(regnum);
@@ -720,24 +714,22 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
 
 	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
 
-	ret = qca8k_mdio_busy_wait(priv, QCA8K_MDIO_MASTER_CTRL,
+	ret = qca8k_mdio_busy_wait(priv->bus, QCA8K_MDIO_MASTER_CTRL,
 				   QCA8K_MDIO_MASTER_BUSY);
 	if (ret)
 		goto exit;
 
 	val = qca8k_mii_read32(priv->bus, 0x10 | r2, r1);
-	val &= QCA8K_MDIO_MASTER_DATA_MASK;
 
 exit:
+	/* even if the busy_wait timeouts try to clear the MASTER_EN */
+	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, 0);
+
 	mutex_unlock(&priv->bus->mdio_lock);
 
 	if (val >= 0)
 		val &= QCA8K_MDIO_MASTER_DATA_MASK;
 
-	/* even if the busy_wait timeouts try to clear the MASTER_EN */
-	qca8k_reg_clear(priv, QCA8K_MDIO_MASTER_CTRL,
-			QCA8K_MDIO_MASTER_EN);
-
 	return val;
 }
 
@@ -746,7 +738,14 @@ qca8k_phy_write(struct dsa_switch *ds, int port, int regnum, u16 data)
 {
 	struct qca8k_priv *priv = ds->priv;
 
-	return qca8k_mdio_write(priv, port, regnum, data);
+	/* Check if the legacy mapping should be used and the
+	 * port is not correctly mapped to the right PHY in the
+	 * devicetree
+	 */
+	if (priv->legacy_phy_port_mapping)
+		port = qca8k_port_to_phy(port) % PHY_MAX_ADDR;
+
+	return qca8k_mdio_write(priv->bus, port, regnum, data);
 }
 
 static int
@@ -755,7 +754,14 @@ qca8k_phy_read(struct dsa_switch *ds, int port, int regnum)
 	struct qca8k_priv *priv = ds->priv;
 	int ret;
 
-	ret = qca8k_mdio_read(priv, port, regnum);
+	/* Check if the legacy mapping should be used and the
+	 * port is not correctly mapped to the right PHY in the
+	 * devicetree
+	 */
+	if (priv->legacy_phy_port_mapping)
+		port = qca8k_port_to_phy(port) % PHY_MAX_ADDR;
+
+	ret = qca8k_mdio_read(priv->bus, port, regnum);
 
 	if (ret < 0)
 		return 0xffff;
@@ -763,11 +769,38 @@ qca8k_phy_read(struct dsa_switch *ds, int port, int regnum)
 	return ret;
 }
 
+static int
+qca8k_mdio_register(struct qca8k_priv *priv, struct device_node *mdio)
+{
+	struct dsa_switch *ds = priv->ds;
+	struct mii_bus *bus;
+
+	bus = devm_mdiobus_alloc(ds->dev);
+
+	if (!bus)
+		return -ENOMEM;
+
+	bus->priv = (void *)priv;
+	bus->name = "qca8k slave mii";
+	bus->read = qca8k_mdio_read;
+	bus->write = qca8k_mdio_write;
+	snprintf(bus->id, MII_BUS_ID_SIZE, "qca8k-%d",
+		 ds->index);
+
+	bus->parent = ds->dev;
+	bus->phy_mask = ~ds->phys_mii_mask;
+
+	ds->slave_mii_bus = bus;
+
+	return devm_of_mdiobus_register(priv->dev, bus, mdio);
+}
+
 static int
 qca8k_setup_mdio_bus(struct qca8k_priv *priv)
 {
 	u32 internal_mdio_mask = 0, external_mdio_mask = 0, reg;
-	struct device_node *ports, *port;
+	struct device_node *ports, *port, *mdio;
+	phy_interface_t mode;
 	int err;
 
 	ports = of_get_child_by_name(priv->dev->of_node, "ports");
@@ -788,7 +821,10 @@ qca8k_setup_mdio_bus(struct qca8k_priv *priv)
 		if (!dsa_is_user_port(priv->ds, reg))
 			continue;
 
-		if (of_property_read_bool(port, "phy-handle"))
+		of_get_phy_mode(port, &mode);
+
+		if (of_property_read_bool(port, "phy-handle") &&
+		    mode != PHY_INTERFACE_MODE_INTERNAL)
 			external_mdio_mask |= BIT(reg);
 		else
 			internal_mdio_mask |= BIT(reg);
@@ -825,8 +861,23 @@ qca8k_setup_mdio_bus(struct qca8k_priv *priv)
 				       QCA8K_MDIO_MASTER_EN);
 	}
 
+	/* Check if the devicetree declare the port:phy mapping */
+	mdio = of_get_child_by_name(priv->dev->of_node, "mdio");
+	if (of_device_is_available(mdio)) {
+		err = qca8k_mdio_register(priv, mdio);
+		if (err)
+			of_node_put(mdio);
+
+		return err;
+	}
+
+	/* If a mapping can't be found the legacy mapping is used,
+	 * using the qca8k_port_to_phy function
+	 */
+	priv->legacy_phy_port_mapping = true;
 	priv->ops.phy_read = qca8k_phy_read;
 	priv->ops.phy_write = qca8k_phy_write;
+
 	return 0;
 }
 
@@ -1212,7 +1263,8 @@ qca8k_phylink_validate(struct dsa_switch *ds, int port,
 	case 5:
 		/* Internal PHY */
 		if (state->interface != PHY_INTERFACE_MODE_NA &&
-		    state->interface != PHY_INTERFACE_MODE_GMII)
+		    state->interface != PHY_INTERFACE_MODE_GMII &&
+		    state->interface != PHY_INTERFACE_MODE_INTERNAL)
 			goto unsupported;
 		break;
 	case 6: /* 2nd CPU port / external PHY */
diff --git a/drivers/net/dsa/qca8k.h b/drivers/net/dsa/qca8k.h
index d365f85ab34f..ed3b05ad6745 100644
--- a/drivers/net/dsa/qca8k.h
+++ b/drivers/net/dsa/qca8k.h
@@ -255,6 +255,7 @@ struct qca8k_priv {
 	u8 switch_revision;
 	u8 rgmii_tx_delay;
 	u8 rgmii_rx_delay;
+	bool legacy_phy_port_mapping;
 	struct regmap *regmap;
 	struct mii_bus *bus;
 	struct ar8xxx_port_status port_sts[QCA8K_NUM_PORTS];
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 21/25] devicetree: bindings: dsa: qca8k: Document internal mdio definition
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (19 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 20/25] net: dsa: qca8k: add support for internal phy and internal mdio Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 22/25] net: dsa: qca8k: improve internal mdio read/write bus access Ansuel Smith
                   ` (4 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith, Rob Herring

Document new way of declare mapping of internal PHY to port.
The new implementation directly declare the PHY connected to the port
by adding a node in the switch node. The driver detect this and register
an internal mdiobus using the mapping defined in the mdio node.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
---
 .../devicetree/bindings/net/dsa/qca8k.txt     | 39 +++++++++++++++++++
 1 file changed, 39 insertions(+)

diff --git a/Documentation/devicetree/bindings/net/dsa/qca8k.txt b/Documentation/devicetree/bindings/net/dsa/qca8k.txt
index 1daf68e7ae19..8c73f67c43ca 100644
--- a/Documentation/devicetree/bindings/net/dsa/qca8k.txt
+++ b/Documentation/devicetree/bindings/net/dsa/qca8k.txt
@@ -21,6 +21,10 @@ described in dsa/dsa.txt. If the QCA8K switch is connect to a SoC's external
 mdio-bus each subnode describing a port needs to have a valid phandle
 referencing the internal PHY it is connected to. This is because there's no
 N:N mapping of port and PHY id.
+To declare the internal mdio-bus configuration, declare a mdio node in the
+switch node and declare the phandle for the port referencing the internal
+PHY is connected to. In this config a internal mdio-bus is registered and
+the mdio MASTER is used as communication.
 
 Don't use mixed external and internal mdio-bus configurations, as this is
 not supported by the hardware.
@@ -150,26 +154,61 @@ for the internal master mdio-bus configuration:
 				port@1 {
 					reg = <1>;
 					label = "lan1";
+					phy-mode = "internal";
+					phy-handle = <&phy_port1>;
 				};
 
 				port@2 {
 					reg = <2>;
 					label = "lan2";
+					phy-mode = "internal";
+					phy-handle = <&phy_port2>;
 				};
 
 				port@3 {
 					reg = <3>;
 					label = "lan3";
+					phy-mode = "internal";
+					phy-handle = <&phy_port3>;
 				};
 
 				port@4 {
 					reg = <4>;
 					label = "lan4";
+					phy-mode = "internal";
+					phy-handle = <&phy_port4>;
 				};
 
 				port@5 {
 					reg = <5>;
 					label = "wan";
+					phy-mode = "internal";
+					phy-handle = <&phy_port5>;
+				};
+			};
+
+			mdio {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				phy_port1: phy@0 {
+					reg = <0>;
+				};
+
+				phy_port2: phy@1 {
+					reg = <1>;
+				};
+
+				phy_port3: phy@2 {
+					reg = <2>;
+				};
+
+				phy_port4: phy@3 {
+					reg = <3>;
+				};
+
+				phy_port5: phy@4 {
+					reg = <4>;
 				};
 			};
 		};
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 22/25] net: dsa: qca8k: improve internal mdio read/write bus access
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (20 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 21/25] devicetree: bindings: dsa: qca8k: Document internal mdio definition Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 23/25] net: dsa: qca8k: pass switch_revision info to phy dev_flags Ansuel Smith
                   ` (3 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Improve the internal mdio read/write bus access by caching the value
without accessing it for every read/write.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/dsa/qca8k.c | 28 +++++++++++++++-------------
 1 file changed, 15 insertions(+), 13 deletions(-)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index ba288181fd1a..ccb3d89cf58c 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -655,6 +655,7 @@ static int
 qca8k_mdio_write(struct mii_bus *salve_bus, int phy, int regnum, u16 data)
 {
 	struct qca8k_priv *priv = salve_bus->priv;
+	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
 	u32 val;
 	int ret;
@@ -669,22 +670,22 @@ qca8k_mdio_write(struct mii_bus *salve_bus, int phy, int regnum, u16 data)
 
 	qca8k_split_addr(QCA8K_MDIO_MASTER_CTRL, &r1, &r2, &page);
 
-	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	ret = qca8k_set_page(priv->bus, page);
+	ret = qca8k_set_page(bus, page);
 	if (ret)
 		goto exit;
 
-	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
+	qca8k_mii_write32(bus, 0x10 | r2, r1, val);
 
-	ret = qca8k_mdio_busy_wait(priv->bus, QCA8K_MDIO_MASTER_CTRL,
+	ret = qca8k_mdio_busy_wait(bus, QCA8K_MDIO_MASTER_CTRL,
 				   QCA8K_MDIO_MASTER_BUSY);
 
 exit:
 	/* even if the busy_wait timeouts try to clear the MASTER_EN */
-	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, 0);
+	qca8k_mii_write32(bus, 0x10 | r2, r1, 0);
 
-	mutex_unlock(&priv->bus->mdio_lock);
+	mutex_unlock(&bus->mdio_lock);
 
 	return ret;
 }
@@ -693,6 +694,7 @@ static int
 qca8k_mdio_read(struct mii_bus *salve_bus, int phy, int regnum)
 {
 	struct qca8k_priv *priv = salve_bus->priv;
+	struct mii_bus *bus = priv->bus;
 	u16 r1, r2, page;
 	u32 val;
 	int ret;
@@ -706,26 +708,26 @@ qca8k_mdio_read(struct mii_bus *salve_bus, int phy, int regnum)
 
 	qca8k_split_addr(QCA8K_MDIO_MASTER_CTRL, &r1, &r2, &page);
 
-	mutex_lock_nested(&priv->bus->mdio_lock, MDIO_MUTEX_NESTED);
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
 
-	ret = qca8k_set_page(priv->bus, page);
+	ret = qca8k_set_page(bus, page);
 	if (ret)
 		goto exit;
 
-	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, val);
+	qca8k_mii_write32(bus, 0x10 | r2, r1, val);
 
-	ret = qca8k_mdio_busy_wait(priv->bus, QCA8K_MDIO_MASTER_CTRL,
+	ret = qca8k_mdio_busy_wait(bus, QCA8K_MDIO_MASTER_CTRL,
 				   QCA8K_MDIO_MASTER_BUSY);
 	if (ret)
 		goto exit;
 
-	val = qca8k_mii_read32(priv->bus, 0x10 | r2, r1);
+	val = qca8k_mii_read32(bus, 0x10 | r2, r1);
 
 exit:
 	/* even if the busy_wait timeouts try to clear the MASTER_EN */
-	qca8k_mii_write32(priv->bus, 0x10 | r2, r1, 0);
+	qca8k_mii_write32(bus, 0x10 | r2, r1, 0);
 
-	mutex_unlock(&priv->bus->mdio_lock);
+	mutex_unlock(&bus->mdio_lock);
 
 	if (val >= 0)
 		val &= QCA8K_MDIO_MASTER_DATA_MASK;
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 23/25] net: dsa: qca8k: pass switch_revision info to phy dev_flags
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (21 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 22/25] net: dsa: qca8k: improve internal mdio read/write bus access Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 24/25] net: phy: at803x: clean whitespace errors Ansuel Smith
                   ` (2 subsequent siblings)
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Define get_phy_flags to pass switch_Revision needed to tweak the
internal PHY with debug values based on the revision.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/net/dsa/qca8k.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/net/dsa/qca8k.c b/drivers/net/dsa/qca8k.c
index ccb3d89cf58c..4753228f02b3 100644
--- a/drivers/net/dsa/qca8k.c
+++ b/drivers/net/dsa/qca8k.c
@@ -1732,6 +1732,22 @@ qca8k_port_vlan_del(struct dsa_switch *ds, int port,
 	return ret;
 }
 
+static u32 qca8k_get_phy_flags(struct dsa_switch *ds, int port)
+{
+	struct qca8k_priv *priv = ds->priv;
+
+	/* Communicate to the phy internal driver the switch revision.
+	 * Based on the switch revision different values needs to be
+	 * set to the dbg and mmd reg on the phy.
+	 * The first 2 bit are used to communicate the switch revision
+	 * to the phy driver.
+	 */
+	if (port > 0 && port < 6)
+		return priv->switch_revision;
+
+	return 0;
+}
+
 static enum dsa_tag_protocol
 qca8k_get_tag_protocol(struct dsa_switch *ds, int port,
 		       enum dsa_tag_protocol mp)
@@ -1765,6 +1781,7 @@ static const struct dsa_switch_ops qca8k_switch_ops = {
 	.phylink_mac_config	= qca8k_phylink_mac_config,
 	.phylink_mac_link_down	= qca8k_phylink_mac_link_down,
 	.phylink_mac_link_up	= qca8k_phylink_mac_link_up,
+	.get_phy_flags		= qca8k_get_phy_flags,
 };
 
 static int qca8k_read_switch_id(struct qca8k_priv *priv)
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 24/25] net: phy: at803x: clean whitespace errors
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (22 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 23/25] net: dsa: qca8k: pass switch_revision info to phy dev_flags Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 21:00 ` [PATCH net-next v6 25/25] net: phy: add support for qca8k switch internal PHY in at803x Ansuel Smith
  2021-05-14 22:40 ` [PATCH net-next v6 00/25] Multiple improvement to qca8k stability patchwork-bot+netdevbpf
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Clean any whitespace errors and fix not aligned define.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/phy/at803x.c | 30 ++++++++++++++++--------------
 1 file changed, 16 insertions(+), 14 deletions(-)

diff --git a/drivers/net/phy/at803x.c b/drivers/net/phy/at803x.c
index 32af52dd5aed..d2378a73de6f 100644
--- a/drivers/net/phy/at803x.c
+++ b/drivers/net/phy/at803x.c
@@ -83,8 +83,8 @@
 #define AT803X_MODE_CFG_MASK			0x0F
 #define AT803X_MODE_CFG_SGMII			0x01
 
-#define AT803X_PSSR			0x11	/*PHY-Specific Status Register*/
-#define AT803X_PSSR_MR_AN_COMPLETE	0x0200
+#define AT803X_PSSR				0x11	/*PHY-Specific Status Register*/
+#define AT803X_PSSR_MR_AN_COMPLETE		0x0200
 
 #define AT803X_DEBUG_REG_0			0x00
 #define AT803X_DEBUG_RX_CLK_DLY_EN		BIT(15)
@@ -128,24 +128,28 @@
 #define AT803X_CLK_OUT_STRENGTH_HALF		1
 #define AT803X_CLK_OUT_STRENGTH_QUARTER		2
 
-#define AT803X_DEFAULT_DOWNSHIFT 5
-#define AT803X_MIN_DOWNSHIFT 2
-#define AT803X_MAX_DOWNSHIFT 9
+#define AT803X_DEFAULT_DOWNSHIFT		5
+#define AT803X_MIN_DOWNSHIFT			2
+#define AT803X_MAX_DOWNSHIFT			9
 
 #define AT803X_MMD3_SMARTEEE_CTL1		0x805b
 #define AT803X_MMD3_SMARTEEE_CTL2		0x805c
 #define AT803X_MMD3_SMARTEEE_CTL3		0x805d
 #define AT803X_MMD3_SMARTEEE_CTL3_LPI_EN	BIT(8)
 
-#define ATH9331_PHY_ID 0x004dd041
-#define ATH8030_PHY_ID 0x004dd076
-#define ATH8031_PHY_ID 0x004dd074
-#define ATH8032_PHY_ID 0x004dd023
-#define ATH8035_PHY_ID 0x004dd072
+#define ATH9331_PHY_ID				0x004dd041
+#define ATH8030_PHY_ID				0x004dd076
+#define ATH8031_PHY_ID				0x004dd074
+#define ATH8032_PHY_ID				0x004dd023
+#define ATH8035_PHY_ID				0x004dd072
 #define AT8030_PHY_ID_MASK			0xffffffef
 
-#define AT803X_PAGE_FIBER		0
-#define AT803X_PAGE_COPPER		1
+#define AT803X_PAGE_FIBER			0
+#define AT803X_PAGE_COPPER			1
+
+/* don't turn off internal PLL */
+#define AT803X_KEEP_PLL_ENABLED			BIT(0)
+#define AT803X_DISABLE_SMARTEEE			BIT(1)
 
 MODULE_DESCRIPTION("Qualcomm Atheros AR803x PHY driver");
 MODULE_AUTHOR("Matus Ujhelyi");
@@ -153,8 +157,6 @@ MODULE_LICENSE("GPL");
 
 struct at803x_priv {
 	int flags;
-#define AT803X_KEEP_PLL_ENABLED	BIT(0)	/* don't turn off internal PLL */
-#define AT803X_DISABLE_SMARTEEE	BIT(1)
 	u16 clk_25m_reg;
 	u16 clk_25m_mask;
 	u8 smarteee_lpi_tw_1g;
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* [PATCH net-next v6 25/25] net: phy: add support for qca8k switch internal PHY in at803x
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (23 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 24/25] net: phy: at803x: clean whitespace errors Ansuel Smith
@ 2021-05-14 21:00 ` Ansuel Smith
  2021-05-14 22:40 ` [PATCH net-next v6 00/25] Multiple improvement to qca8k stability patchwork-bot+netdevbpf
  25 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 21:00 UTC (permalink / raw)
  To: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	Russell King, netdev, devicetree, linux-kernel
  Cc: Ansuel Smith

Since the at803x share the same regs, it's assumed they are based on the
same implementation. Make it part of the at803x PHY driver to skip
having redudant code.
Add initial support for qca8k internal PHYs. The internal PHYs requires
special mmd and debug values to be set based on the switch revision
passwd using the dev_flags. Supports output of idle, receive and eee_wake
errors stats.
Some debug values sets can't be translated as the documentation lacks any
reference about them.

Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
---
 drivers/net/phy/Kconfig  |   5 +-
 drivers/net/phy/at803x.c | 132 ++++++++++++++++++++++++++++++++++++++-
 2 files changed, 134 insertions(+), 3 deletions(-)

diff --git a/drivers/net/phy/Kconfig b/drivers/net/phy/Kconfig
index 288bf405ebdb..25511f39b01f 100644
--- a/drivers/net/phy/Kconfig
+++ b/drivers/net/phy/Kconfig
@@ -247,10 +247,11 @@ config NXP_TJA11XX_PHY
 	  Currently supports the NXP TJA1100 and TJA1101 PHY.
 
 config AT803X_PHY
-	tristate "Qualcomm Atheros AR803X PHYs"
+	tristate "Qualcomm Atheros AR803X PHYs and QCA833x PHYs"
 	depends on REGULATOR
 	help
-	  Currently supports the AR8030, AR8031, AR8033 and AR8035 model
+	  Currently supports the AR8030, AR8031, AR8033, AR8035 and internal
+	  QCA8337(Internal qca8k PHY) model
 
 config QSEMI_PHY
 	tristate "Quality Semiconductor PHYs"
diff --git a/drivers/net/phy/at803x.c b/drivers/net/phy/at803x.c
index d2378a73de6f..6697c9368b40 100644
--- a/drivers/net/phy/at803x.c
+++ b/drivers/net/phy/at803x.c
@@ -92,10 +92,16 @@
 #define AT803X_DEBUG_REG_5			0x05
 #define AT803X_DEBUG_TX_CLK_DLY_EN		BIT(8)
 
+#define AT803X_DEBUG_REG_3C			0x3C
+
+#define AT803X_DEBUG_REG_3D			0x3D
+
 #define AT803X_DEBUG_REG_1F			0x1F
 #define AT803X_DEBUG_PLL_ON			BIT(2)
 #define AT803X_DEBUG_RGMII_1V8			BIT(3)
 
+#define MDIO_AZ_DEBUG				0x800D
+
 /* AT803x supports either the XTAL input pad, an internal PLL or the
  * DSP as clock reference for the clock output pad. The XTAL reference
  * is only used for 25 MHz output, all other frequencies need the PLL.
@@ -144,6 +150,12 @@
 #define ATH8035_PHY_ID				0x004dd072
 #define AT8030_PHY_ID_MASK			0xffffffef
 
+#define QCA8327_PHY_ID				0x004dd034
+#define QCA8337_PHY_ID				0x004dd036
+#define QCA8K_PHY_ID_MASK			0xffffffff
+
+#define QCA8K_DEVFLAGS_REVISION_MASK		GENMASK(2, 0)
+
 #define AT803X_PAGE_FIBER			0
 #define AT803X_PAGE_COPPER			1
 
@@ -155,6 +167,24 @@ MODULE_DESCRIPTION("Qualcomm Atheros AR803x PHY driver");
 MODULE_AUTHOR("Matus Ujhelyi");
 MODULE_LICENSE("GPL");
 
+enum stat_access_type {
+	PHY,
+	MMD
+};
+
+struct at803x_hw_stat {
+	const char *string;
+	u8 reg;
+	u32 mask;
+	enum stat_access_type access_type;
+};
+
+static struct at803x_hw_stat at803x_hw_stats[] = {
+	{ "phy_idle_errors", 0xa, GENMASK(7, 0), PHY},
+	{ "phy_receive_errors", 0x15, GENMASK(15, 0), PHY},
+	{ "eee_wake_errors", 0x16, GENMASK(15, 0), MMD},
+};
+
 struct at803x_priv {
 	int flags;
 	u16 clk_25m_reg;
@@ -164,6 +194,7 @@ struct at803x_priv {
 	struct regulator_dev *vddio_rdev;
 	struct regulator_dev *vddh_rdev;
 	struct regulator *vddio;
+	u64 stats[ARRAY_SIZE(at803x_hw_stats)];
 };
 
 struct at803x_context {
@@ -175,6 +206,17 @@ struct at803x_context {
 	u16 led_control;
 };
 
+static int at803x_debug_reg_write(struct phy_device *phydev, u16 reg, u16 data)
+{
+	int ret;
+
+	ret = phy_write(phydev, AT803X_DEBUG_ADDR, reg);
+	if (ret < 0)
+		return ret;
+
+	return phy_write(phydev, AT803X_DEBUG_DATA, data);
+}
+
 static int at803x_debug_reg_read(struct phy_device *phydev, u16 reg)
 {
 	int ret;
@@ -337,6 +379,53 @@ static void at803x_get_wol(struct phy_device *phydev,
 		wol->wolopts |= WAKE_MAGIC;
 }
 
+static int at803x_get_sset_count(struct phy_device *phydev)
+{
+	return ARRAY_SIZE(at803x_hw_stats);
+}
+
+static void at803x_get_strings(struct phy_device *phydev, u8 *data)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(at803x_hw_stats); i++) {
+		strscpy(data + i * ETH_GSTRING_LEN,
+			at803x_hw_stats[i].string, ETH_GSTRING_LEN);
+	}
+}
+
+static u64 at803x_get_stat(struct phy_device *phydev, int i)
+{
+	struct at803x_hw_stat stat = at803x_hw_stats[i];
+	struct at803x_priv *priv = phydev->priv;
+	int val;
+	u64 ret;
+
+	if (stat.access_type == MMD)
+		val = phy_read_mmd(phydev, MDIO_MMD_PCS, stat.reg);
+	else
+		val = phy_read(phydev, stat.reg);
+
+	if (val < 0) {
+		ret = U64_MAX;
+	} else {
+		val = val & stat.mask;
+		priv->stats[i] += val;
+		ret = priv->stats[i];
+	}
+
+	return ret;
+}
+
+static void at803x_get_stats(struct phy_device *phydev,
+			     struct ethtool_stats *stats, u64 *data)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(at803x_hw_stats); i++)
+		data[i] = at803x_get_stat(phydev, i);
+}
+
 static int at803x_suspend(struct phy_device *phydev)
 {
 	int value;
@@ -1172,6 +1261,34 @@ static int at803x_cable_test_start(struct phy_device *phydev)
 	return 0;
 }
 
+static int qca83xx_config_init(struct phy_device *phydev)
+{
+	u8 switch_revision;
+
+	switch_revision = phydev->dev_flags & QCA8K_DEVFLAGS_REVISION_MASK;
+
+	switch (switch_revision) {
+	case 1:
+		/* For 100M waveform */
+		at803x_debug_reg_write(phydev, AT803X_DEBUG_REG_0, 0x02ea);
+		/* Turn on Gigabit clock */
+		at803x_debug_reg_write(phydev, AT803X_DEBUG_REG_3D, 0x68a0);
+		break;
+
+	case 2:
+		phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_EEE_ADV, 0x0);
+		fallthrough;
+	case 4:
+		phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_AZ_DEBUG, 0x803f);
+		at803x_debug_reg_write(phydev, AT803X_DEBUG_REG_3D, 0x6860);
+		at803x_debug_reg_write(phydev, AT803X_DEBUG_REG_5, 0x2c46);
+		at803x_debug_reg_write(phydev, AT803X_DEBUG_REG_3C, 0x6000);
+		break;
+	}
+
+	return 0;
+}
+
 static struct phy_driver at803x_driver[] = {
 {
 	/* Qualcomm Atheros AR8035 */
@@ -1268,7 +1385,20 @@ static struct phy_driver at803x_driver[] = {
 	.read_status		= at803x_read_status,
 	.soft_reset		= genphy_soft_reset,
 	.config_aneg		= at803x_config_aneg,
-} };
+}, {
+	/* QCA8337 */
+	.phy_id = QCA8337_PHY_ID,
+	.phy_id_mask = QCA8K_PHY_ID_MASK,
+	.name = "QCA PHY 8337",
+	/* PHY_GBIT_FEATURES */
+	.probe = at803x_probe,
+	.flags = PHY_IS_INTERNAL,
+	.config_init = qca83xx_config_init,
+	.soft_reset = genphy_soft_reset,
+	.get_sset_count = at803x_get_sset_count,
+	.get_strings = at803x_get_strings,
+	.get_stats = at803x_get_stats,
+}, };
 
 module_phy_driver(at803x_driver);
 
-- 
2.30.2


^ permalink raw reply related	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 00/25] Multiple improvement to qca8k stability
  2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
                   ` (24 preceding siblings ...)
  2021-05-14 21:00 ` [PATCH net-next v6 25/25] net: phy: add support for qca8k switch internal PHY in at803x Ansuel Smith
@ 2021-05-14 22:40 ` patchwork-bot+netdevbpf
  25 siblings, 0 replies; 33+ messages in thread
From: patchwork-bot+netdevbpf @ 2021-05-14 22:40 UTC (permalink / raw)
  To: Ansuel Smith
  Cc: andrew, vivien.didelot, f.fainelli, olteanv, davem, kuba,
	robh+dt, hkallweit1, linux, netdev, devicetree, linux-kernel

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

On Fri, 14 May 2021 22:59:50 +0200 you wrote:
> Currently qca8337 switch are widely used on ipq8064 based router.
> On these particular router it was notice a very unstable switch with
> port not link detected as link with unknown speed, port dropping
> randomly and general unreliability. Lots of testing and comparison
> between this dsa driver and the original qsdk driver showed lack of some
> additional delay and values. A main difference arised from the original
> driver and the dsa one. The original driver didn't use MASTER regs to
> read phy status and the dedicated mdio driver worked correctly. Now that
> the dsa driver actually use these regs, it was found that these special
> read/write operation required mutual exclusion to normal
> qca8k_read/write operation. The add of mutex for these operation fixed
> the random port dropping and now only the actual linked port randomly
> dropped. Adding additional delay for set_page operation and fixing a bug
> in the mdio dedicated driver fixed also this problem. The current driver
> requires also more time to apply vlan switch. All of these changes and
> tweak permit a now very stable and reliable dsa driver and 0 port
> dropping. This series is currently tested by at least 5 user with
> different routers and all reports positive results and no problems.
> 
> [...]

Here is the summary with links:
  - [net-next,v6,01/25] net: dsa: qca8k: change simple print to dev variant
    https://git.kernel.org/netdev/net-next/c/5d9e068402dc
  - [net-next,v6,02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait
    https://git.kernel.org/netdev/net-next/c/2ad255f2faaf
  - [net-next,v6,03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access
    https://git.kernel.org/netdev/net-next/c/504bf6593182
  - [net-next,v6,04/25] net: dsa: qca8k: handle qca8k_set_page errors
    https://git.kernel.org/netdev/net-next/c/ba5707ec58cf
  - [net-next,v6,05/25] net: dsa: qca8k: handle error with qca8k_read operation
    https://git.kernel.org/netdev/net-next/c/028f5f8ef44f
  - [net-next,v6,06/25] net: dsa: qca8k: handle error with qca8k_write operation
    https://git.kernel.org/netdev/net-next/c/d7805757c75c
  - [net-next,v6,07/25] net: dsa: qca8k: handle error with qca8k_rmw operation
    https://git.kernel.org/netdev/net-next/c/aaf421425cbd
  - [net-next,v6,08/25] net: dsa: qca8k: handle error from qca8k_busy_wait
    https://git.kernel.org/netdev/net-next/c/b7c818d19492
  - [net-next,v6,09/25] net: dsa: qca8k: add support for qca8327 switch
    https://git.kernel.org/netdev/net-next/c/6e82a457e062
  - [net-next,v6,10/25] devicetree: net: dsa: qca8k: Document new compatible qca8327
    https://git.kernel.org/netdev/net-next/c/227a9ffc1bc7
  - [net-next,v6,11/25] net: dsa: qca8k: add priority tweak to qca8337 switch
    https://git.kernel.org/netdev/net-next/c/83a3ceb39b24
  - [net-next,v6,12/25] net: dsa: qca8k: limit port5 delay to qca8337
    https://git.kernel.org/netdev/net-next/c/5bf9ff3b9fb5
  - [net-next,v6,13/25] net: dsa: qca8k: add GLOBAL_FC settings needed for qca8327
    https://git.kernel.org/netdev/net-next/c/0fc57e4b5e39
  - [net-next,v6,14/25] net: dsa: qca8k: add support for switch rev
    https://git.kernel.org/netdev/net-next/c/95ffeaf18b3b
  - [net-next,v6,15/25] net: dsa: qca8k: add ethernet-ports fallback to setup_mdio_bus
    https://git.kernel.org/netdev/net-next/c/1ee0591a1093
  - [net-next,v6,16/25] net: dsa: qca8k: make rgmii delay configurable
    https://git.kernel.org/netdev/net-next/c/e4b9977cee15
  - [net-next,v6,17/25] net: dsa: qca8k: clear MASTER_EN after phy read/write
    https://git.kernel.org/netdev/net-next/c/63c33bbfeb68
  - [net-next,v6,18/25] net: dsa: qca8k: dsa: qca8k: protect MASTER busy_wait with mdio mutex
    https://git.kernel.org/netdev/net-next/c/60df02b6ea45
  - [net-next,v6,19/25] net: dsa: qca8k: enlarge mdio delay and timeout
    https://git.kernel.org/netdev/net-next/c/617960d72e93
  - [net-next,v6,20/25] net: dsa: qca8k: add support for internal phy and internal mdio
    https://git.kernel.org/netdev/net-next/c/759bafb8a322
  - [net-next,v6,21/25] devicetree: bindings: dsa: qca8k: Document internal mdio definition
    https://git.kernel.org/netdev/net-next/c/0c994a28e751
  - [net-next,v6,22/25] net: dsa: qca8k: improve internal mdio read/write bus access
    https://git.kernel.org/netdev/net-next/c/b7ebac354d54
  - [net-next,v6,23/25] net: dsa: qca8k: pass switch_revision info to phy dev_flags
    https://git.kernel.org/netdev/net-next/c/a46aec02bc06
  - [net-next,v6,24/25] net: phy: at803x: clean whitespace errors
    https://git.kernel.org/netdev/net-next/c/d0e13fd5626c
  - [net-next,v6,25/25] net: phy: add support for qca8k switch internal PHY in at803x
    https://git.kernel.org/netdev/net-next/c/272833b9b3b3

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html



^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant
  2021-05-14 20:59 ` [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant Ansuel Smith
@ 2021-05-14 22:43   ` Russell King (Oracle)
  0 siblings, 0 replies; 33+ messages in thread
From: Russell King (Oracle) @ 2021-05-14 22:43 UTC (permalink / raw)
  To: Ansuel Smith
  Cc: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	netdev, devicetree, linux-kernel

On Fri, May 14, 2021 at 10:59:51PM +0200, Ansuel Smith wrote:
> Change pr_err and pr_warn to dev variant.
> 
> Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
> Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>

Thanks!

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait
  2021-05-14 20:59 ` [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait Ansuel Smith
@ 2021-05-14 22:52   ` Russell King (Oracle)
  2021-05-14 23:07     ` Ansuel Smith
  0 siblings, 1 reply; 33+ messages in thread
From: Russell King (Oracle) @ 2021-05-14 22:52 UTC (permalink / raw)
  To: Ansuel Smith
  Cc: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	netdev, devicetree, linux-kernel

On Fri, May 14, 2021 at 10:59:52PM +0200, Ansuel Smith wrote:
> Use iopoll macro instead of while loop.
> 
> Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

This doesn't look quite right to me.

>  static int
>  qca8k_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
>  {
> -	unsigned long timeout;
> -
> -	timeout = jiffies + msecs_to_jiffies(20);
> +	u32 val;

val is unsigned here.

> +	/* Check if qca8k_read has failed for a different reason
> +	 * before returning -ETIMEDOUT
> +	 */
> +	if (ret < 0 && val < 0)

but here you are checking it for a negative number - this will always be
false, making the conditional code unreachable. Either the test is wrong,
or the type of val is wrong. Please resolve.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access
  2021-05-14 20:59 ` [PATCH net-next v6 03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access Ansuel Smith
@ 2021-05-14 22:53   ` Russell King (Oracle)
  0 siblings, 0 replies; 33+ messages in thread
From: Russell King (Oracle) @ 2021-05-14 22:53 UTC (permalink / raw)
  To: Ansuel Smith
  Cc: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	netdev, devicetree, linux-kernel

On Fri, May 14, 2021 at 10:59:53PM +0200, Ansuel Smith wrote:
> Put bus in local variable to improve faster access to the mdio bus.
> 
> Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 04/25] net: dsa: qca8k: handle qca8k_set_page errors
  2021-05-14 20:59 ` [PATCH net-next v6 04/25] net: dsa: qca8k: handle qca8k_set_page errors Ansuel Smith
@ 2021-05-14 22:55   ` Russell King (Oracle)
  0 siblings, 0 replies; 33+ messages in thread
From: Russell King (Oracle) @ 2021-05-14 22:55 UTC (permalink / raw)
  To: Ansuel Smith
  Cc: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	netdev, devicetree, linux-kernel

On Fri, May 14, 2021 at 10:59:54PM +0200, Ansuel Smith wrote:
> With a remote possibility, the set_page function can fail. Since this is
> a critical part of the write/read qca8k regs, propagate the error and
> terminate any read/write operation.
> 
> Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>

Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 05/25] net: dsa: qca8k: handle error with qca8k_read operation
  2021-05-14 20:59 ` [PATCH net-next v6 05/25] net: dsa: qca8k: handle error with qca8k_read operation Ansuel Smith
@ 2021-05-14 23:01   ` Russell King (Oracle)
  0 siblings, 0 replies; 33+ messages in thread
From: Russell King (Oracle) @ 2021-05-14 23:01 UTC (permalink / raw)
  To: Ansuel Smith
  Cc: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	netdev, devicetree, linux-kernel

On Fri, May 14, 2021 at 10:59:55PM +0200, Ansuel Smith wrote:
> -static void
> +static int
>  qca8k_fdb_read(struct qca8k_priv *priv, struct qca8k_fdb *fdb)
>  {
> -	u32 reg[4];
> +	u32 reg[4], val;

val is unsigned.

>  	int i;
>  
>  	/* load the ARL table into an array */
> -	for (i = 0; i < 4; i++)
> -		reg[i] = qca8k_read(priv, QCA8K_REG_ATU_DATA0 + (i * 4));
> +	for (i = 0; i < 4; i++) {
> +		val = qca8k_read(priv, QCA8K_REG_ATU_DATA0 + (i * 4));
> +		if (val < 0)
> +			return val;

So this return statement will never be reached.

> @@ -374,6 +386,8 @@ qca8k_fdb_access(struct qca8k_priv *priv, enum qca8k_fdb_cmd cmd, int port)
>  	/* Check for table full violation when adding an entry */
>  	if (cmd == QCA8K_FDB_LOAD) {
>  		reg = qca8k_read(priv, QCA8K_REG_ATU_FUNC);
> +		if (reg < 0)
> +			return reg;

"reg" here is also a u32, and therefore unsigned, so this will have no
effect.

>  		if (reg & QCA8K_ATU_FUNC_FULL)
>  			return -1;
>  	}
> @@ -388,10 +402,10 @@ qca8k_fdb_next(struct qca8k_priv *priv, struct qca8k_fdb *fdb, int port)
>  
>  	qca8k_fdb_write(priv, fdb->vid, fdb->port_mask, fdb->mac, fdb->aging);
>  	ret = qca8k_fdb_access(priv, QCA8K_FDB_NEXT, port);
> -	if (ret >= 0)
> -		qca8k_fdb_read(priv, fdb);
> +	if (ret < 0)
> +		return ret;

This looks fine to me.

>  
> -	return ret;
> +	return qca8k_fdb_read(priv, fdb);
>  }
>  
>  static int
> @@ -449,6 +463,8 @@ qca8k_vlan_access(struct qca8k_priv *priv, enum qca8k_vlan_cmd cmd, u16 vid)
>  	/* Check for table full violation when adding an entry */
>  	if (cmd == QCA8K_VLAN_LOAD) {
>  		reg = qca8k_read(priv, QCA8K_REG_VTU_FUNC1);
> +		if (reg < 0)
> +			return reg;

reg is unsigned... unreachable.

>  		if (reg & QCA8K_VTU_FUNC1_FULL)
>  			return -ENOMEM;
>  	}
> @@ -475,6 +491,8 @@ qca8k_vlan_add(struct qca8k_priv *priv, u8 port, u16 vid, bool untagged)
>  		goto out;
>  
>  	reg = qca8k_read(priv, QCA8K_REG_VTU_FUNC0);
> +	if (reg < 0)
> +		return reg;

reg is unsigned... unreachable.

>  	reg |= QCA8K_VTU_FUNC0_VALID | QCA8K_VTU_FUNC0_IVL_EN;
>  	reg &= ~(QCA8K_VTU_FUNC0_EG_MODE_MASK << QCA8K_VTU_FUNC0_EG_MODE_S(port));
>  	if (untagged)
> @@ -506,6 +524,8 @@ qca8k_vlan_del(struct qca8k_priv *priv, u8 port, u16 vid)
>  		goto out;
>  
>  	reg = qca8k_read(priv, QCA8K_REG_VTU_FUNC0);
> +	if (reg < 0)
> +		return reg;

reg is unsigned... unreachable.

>  	reg &= ~(3 << QCA8K_VTU_FUNC0_EG_MODE_S(port));
>  	reg |= QCA8K_VTU_FUNC0_EG_MODE_NOT <<
>  			QCA8K_VTU_FUNC0_EG_MODE_S(port);
> @@ -621,8 +641,11 @@ qca8k_mdio_read(struct qca8k_priv *priv, int port, u32 regnum)
>  			    QCA8K_MDIO_MASTER_BUSY))
>  		return -ETIMEDOUT;
>  
> -	val = (qca8k_read(priv, QCA8K_MDIO_MASTER_CTRL) &
> -		QCA8K_MDIO_MASTER_DATA_MASK);
> +	val = qca8k_read(priv, QCA8K_MDIO_MASTER_CTRL);
> +	if (val < 0)
> +		return val;

val is unsigned... unreachable.

> +
> +	val &= QCA8K_MDIO_MASTER_DATA_MASK;
>  
>  	return val;
>  }
> @@ -978,6 +1001,8 @@ qca8k_phylink_mac_link_state(struct dsa_switch *ds, int port,
>  	u32 reg;
>  
>  	reg = qca8k_read(priv, QCA8K_REG_PORT_STATUS(port));
> +	if (reg < 0)
> +		return reg;

reg is unsigned... unreachable.

>  
>  	state->link = !!(reg & QCA8K_PORT_STATUS_LINK_UP);
>  	state->an_complete = state->link;
> @@ -1078,18 +1103,26 @@ qca8k_get_ethtool_stats(struct dsa_switch *ds, int port,
>  {
>  	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
>  	const struct qca8k_mib_desc *mib;
> -	u32 reg, i;
> +	u32 reg, i, val;
>  	u64 hi;
>  
>  	for (i = 0; i < ARRAY_SIZE(ar8327_mib); i++) {
>  		mib = &ar8327_mib[i];
>  		reg = QCA8K_PORT_MIB_COUNTER(port) + mib->offset;
>  
> -		data[i] = qca8k_read(priv, reg);
> +		val = qca8k_read(priv, reg);
> +		if (val < 0)
> +			continue;

val is unsigned... unreachable....

> +
>  		if (mib->size == 2) {
>  			hi = qca8k_read(priv, reg + 4);
> -			data[i] |= hi << 32;
> +			if (hi < 0)
> +				continue;

hi is a u64, so this condition is always false.

>  		}
> +
> +		data[i] = val;
> +		if (mib->size == 2)
> +			data[i] |= hi << 32;
>  	}
>  }
>  
> @@ -1107,18 +1140,25 @@ qca8k_set_mac_eee(struct dsa_switch *ds, int port, struct ethtool_eee *eee)
>  {
>  	struct qca8k_priv *priv = (struct qca8k_priv *)ds->priv;
>  	u32 lpi_en = QCA8K_REG_EEE_CTRL_LPI_EN(port);
> +	int ret = 0;

No need to zero-initialise this.

>  	u32 reg;
>  
>  	mutex_lock(&priv->reg_mutex);
>  	reg = qca8k_read(priv, QCA8K_REG_EEE_CTRL);
> +	if (reg < 0) {
> +		ret = reg;
> +		goto exit;
> +	}
> +
>  	if (eee->eee_enabled)
>  		reg |= lpi_en;
>  	else
>  		reg &= ~lpi_en;
>  	qca8k_write(priv, QCA8K_REG_EEE_CTRL, reg);
> -	mutex_unlock(&priv->reg_mutex);
>  
> -	return 0;
> +exit:
> +	mutex_unlock(&priv->reg_mutex);
> +	return ret;
>  }
>  
>  static int
> @@ -1443,6 +1483,9 @@ qca8k_sw_probe(struct mdio_device *mdiodev)
>  
>  	/* read the switches ID register */
>  	id = qca8k_read(priv, QCA8K_REG_MASK_CTRL);
> +	if (id < 0)
> +		return id;

id is unsigned ...

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

^ permalink raw reply	[flat|nested] 33+ messages in thread

* Re: [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait
  2021-05-14 22:52   ` Russell King (Oracle)
@ 2021-05-14 23:07     ` Ansuel Smith
  0 siblings, 0 replies; 33+ messages in thread
From: Ansuel Smith @ 2021-05-14 23:07 UTC (permalink / raw)
  To: Russell King (Oracle)
  Cc: Andrew Lunn, Vivien Didelot, Florian Fainelli, Vladimir Oltean,
	David S. Miller, Jakub Kicinski, Rob Herring, Heiner Kallweit,
	netdev, devicetree, linux-kernel

On Fri, May 14, 2021 at 11:52:25PM +0100, Russell King (Oracle) wrote:
> On Fri, May 14, 2021 at 10:59:52PM +0200, Ansuel Smith wrote:
> > Use iopoll macro instead of while loop.
> > 
> > Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
> > Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> 
> This doesn't look quite right to me.
> 
> >  static int
> >  qca8k_busy_wait(struct qca8k_priv *priv, u32 reg, u32 mask)
> >  {
> > -	unsigned long timeout;
> > -
> > -	timeout = jiffies + msecs_to_jiffies(20);
> > +	u32 val;
> 
> val is unsigned here.
> 
> > +	/* Check if qca8k_read has failed for a different reason
> > +	 * before returning -ETIMEDOUT
> > +	 */
> > +	if (ret < 0 && val < 0)
> 
> but here you are checking it for a negative number - this will always be
> false, making the conditional code unreachable. Either the test is wrong,
> or the type of val is wrong. Please resolve.
>

I know this is wrong and I will fix.
Anyway I tested this and I checked if with u32 a negative value was
actually provided and to my surprise the value was correctly returned.
Any idea why?

> -- 
> RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
> FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

^ permalink raw reply	[flat|nested] 33+ messages in thread

end of thread, other threads:[~2021-05-14 23:07 UTC | newest]

Thread overview: 33+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-05-14 20:59 [PATCH net-next v6 00/25] Multiple improvement to qca8k stability Ansuel Smith
2021-05-14 20:59 ` [PATCH net-next v6 01/25] net: dsa: qca8k: change simple print to dev variant Ansuel Smith
2021-05-14 22:43   ` Russell King (Oracle)
2021-05-14 20:59 ` [PATCH net-next v6 02/25] net: dsa: qca8k: use iopoll macro for qca8k_busy_wait Ansuel Smith
2021-05-14 22:52   ` Russell King (Oracle)
2021-05-14 23:07     ` Ansuel Smith
2021-05-14 20:59 ` [PATCH net-next v6 03/25] net: dsa: qca8k: improve qca8k read/write/rmw bus access Ansuel Smith
2021-05-14 22:53   ` Russell King (Oracle)
2021-05-14 20:59 ` [PATCH net-next v6 04/25] net: dsa: qca8k: handle qca8k_set_page errors Ansuel Smith
2021-05-14 22:55   ` Russell King (Oracle)
2021-05-14 20:59 ` [PATCH net-next v6 05/25] net: dsa: qca8k: handle error with qca8k_read operation Ansuel Smith
2021-05-14 23:01   ` Russell King (Oracle)
2021-05-14 20:59 ` [PATCH net-next v6 06/25] net: dsa: qca8k: handle error with qca8k_write operation Ansuel Smith
2021-05-14 20:59 ` [PATCH net-next v6 07/25] net: dsa: qca8k: handle error with qca8k_rmw operation Ansuel Smith
2021-05-14 20:59 ` [PATCH net-next v6 08/25] net: dsa: qca8k: handle error from qca8k_busy_wait Ansuel Smith
2021-05-14 20:59 ` [PATCH net-next v6 09/25] net: dsa: qca8k: add support for qca8327 switch Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 10/25] devicetree: net: dsa: qca8k: Document new compatible qca8327 Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 11/25] net: dsa: qca8k: add priority tweak to qca8337 switch Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 12/25] net: dsa: qca8k: limit port5 delay to qca8337 Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 13/25] net: dsa: qca8k: add GLOBAL_FC settings needed for qca8327 Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 14/25] net: dsa: qca8k: add support for switch rev Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 15/25] net: dsa: qca8k: add ethernet-ports fallback to setup_mdio_bus Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 16/25] net: dsa: qca8k: make rgmii delay configurable Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 17/25] net: dsa: qca8k: clear MASTER_EN after phy read/write Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 18/25] net: dsa: qca8k: dsa: qca8k: protect MASTER busy_wait with mdio mutex Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 19/25] net: dsa: qca8k: enlarge mdio delay and timeout Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 20/25] net: dsa: qca8k: add support for internal phy and internal mdio Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 21/25] devicetree: bindings: dsa: qca8k: Document internal mdio definition Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 22/25] net: dsa: qca8k: improve internal mdio read/write bus access Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 23/25] net: dsa: qca8k: pass switch_revision info to phy dev_flags Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 24/25] net: phy: at803x: clean whitespace errors Ansuel Smith
2021-05-14 21:00 ` [PATCH net-next v6 25/25] net: phy: add support for qca8k switch internal PHY in at803x Ansuel Smith
2021-05-14 22:40 ` [PATCH net-next v6 00/25] Multiple improvement to qca8k stability patchwork-bot+netdevbpf

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).