[PATCH v3 phy-next 0/8] RCW override for 10G Lynx dynamic protocol reconfiguration

[PATCH v3 phy-next 0/8] RCW override for 10G Lynx dynamic protocol reconfiguration

[Vladimir Oltean]

Previous set "New Generic PHY driver for Lynx 10G SerDes":
https://lore.kernel.org/linux-phy/20260610151952.2141019-1-vladimir.oltean@nxp.com/
introduced the 10G Lynx SerDes driver with a reduced functionality set.
Namely, only minor protocol changes are supported (1GbE <-> 2.5GbE).
The major protocol changes need a procedure named RCW override,
explained in more detail in commits 6/8 and 7/8.

This series adds kernel and device tree binding support for RCW
override, completing the SerDes PHY driver functionality.

Two components are involved:
- drivers/soc/fsl/guts.c (binding is fsl,layerscape-dcfg.yaml) - Device
  Configuration Unit, this is API provider for the SerDes driver to
  request RCW override depending on SoC
- drivers/phy/freescale/phy-fsl-lynx-10g.c - SerDes PHY driver, this is
  API consumer

The guts driver probes on DCFG blocks from multiple Freescale SoC
generations:
- MPC85xx, BSC and QorIQ (PowerPC) are all covered by the
  Documentation/devicetree/bindings/soc/fsl/guts.txt schema
- Layerscape (Arm) is covered by
  Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml

It is ultimately the same hardware block, just that (from what I can
tell) the Layerscape nodes are also compatible with syscon, and PowerPC
aren't.

RCW override has only been validated on select Layerscape SoCs, so
converting guts.txt to a PowerPC schema is out of scope for this
series - we don't even touch that (just in case it gets asked).

Using syscon to map the DCFG_DCSR register block in the Lynx SerDes
driver instead of creating this guts <-> lynx API was considered, but
because the RCW procedure is SoC-specific, it was ruled out for
polluting the SerDes driver. The guts driver is all about SoC awareness
anyway, and it offers some abstraction of all the gory details.

Changes since v2:
- fix error handling in fsl_guts_init() and make sure that all newly
  introduced RCW override API functions fail if fsl_guts_init() failed.
- replace __bf_shf() use with __ffs()
- use read_poll_timeout_atomic() in fsl_guts_rcw_rmw() to clarify that
  DCFG_DCSR writes are supposed to reflect back in DCFG_CCSR
- only operate on lanes on which fsl_guts_lane_init() was called in
  ls2088a_serdes_init_rcwcr()
- put parentheses around (lane) in LS1088A_RCWSR29_SRDS_PRTCL_S1_LNn()
  and LS1088A_RCWSR30_SRDS_PRTCL_S2_LNn() macro expressions

v2 at:
https://lore.kernel.org/linux-phy/20260612210859.266759-8-vladimir.oltean@nxp.com/

Changes since v1:
- add Conor's review tag on 6/8
- update email addresses of DT maintainers
- drop DT maintainers from explicit CC on patch 7/8
- keep devicetree@vger.kernel.org CCed on entire series
- include missing <linux/bitfield.h> in patch 7/8
- namespace SRDS_PRTCL values for LS1046A and LS1088A, even if they are
  the same. For LS1028A (not covered here) they are not.
- prefix SRDS_CLK_SEL_{GMII,XGMII} with LS2088A_
- reorder alphanumerically (LS1046A should come before LS1088A)

Change logs also in individual patches.

v1 at:
https://lore.kernel.org/linux-phy/20260611193940.44416-1-vladimir.oltean@nxp.com/

Cc: Conor Dooley <conor+dt@kernel.org>
Cc: Krzysztof Kozlowski <krzk+dt@kernel.org>
Cc: Rob Herring <robh@kernel.org>

Ioana Ciornei (5):
  soc: fsl: guts: use a macro to encode the DCFG CCSR space
  soc: fsl: guts: add a global structure to hold state
  soc: fsl: guts: add a central fsl_guts_read() function
  soc: fsl: guts: make it easier to determine on which SoC we are
    running
  soc: fsl: guts: implement the RCW override procedure

Vladimir Oltean (3):
  soc: fsl: guts: make fsl_soc_data available after fsl_guts_init()
  dt-bindings: fsl: layerscape-dcfg: define DCFG_DCSR region
  phy: lynx-10g: use RCW override procedure for dynamic protocol change

 .../bindings/soc/fsl/fsl,layerscape-dcfg.yaml |  15 +-
 drivers/phy/freescale/Kconfig                 |   1 +
 drivers/phy/freescale/phy-fsl-lynx-10g.c      |  24 +-
 drivers/soc/fsl/guts.c                        | 438 ++++++++++++++++--
 include/linux/fsl/guts.h                      |  20 +-
 5 files changed, 459 insertions(+), 39 deletions(-)

-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 1/8] soc: fsl: guts: use a macro to encode the DCFG CCSR space

[Vladimir Oltean]

From: Ioana Ciornei <ioana.ciornei@nxp.com>

Instead of using a hardcoded value when iomapping the DCFG CCSR space,
add a new macro for it. The code will be easier to follow this way,
especially when we add support for the DCFG DCSR space as well.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v1->v3: none
---
 drivers/soc/fsl/guts.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
index 9bee7baec2b9..f87ee47c1503 100644
--- a/drivers/soc/fsl/guts.c
+++ b/drivers/soc/fsl/guts.c
@@ -14,6 +14,8 @@
 #include <linux/platform_device.h>
 #include <linux/fsl/guts.h>
 
+#define DCFG_CCSR	0
+
 struct fsl_soc_die_attr {
 	char	*die;
 	u32	svr;
@@ -197,7 +199,7 @@ static int __init fsl_guts_init(void)
 		return 0;
 	soc_data = match->data;
 
-	regs = of_iomap(np, 0);
+	regs = of_iomap(np, DCFG_CCSR);
 	if (!regs) {
 		of_node_put(np);
 		return -ENOMEM;
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state

[Vladimir Oltean]

From: Ioana Ciornei <ioana.ciornei@nxp.com>

Add the fsl_soc_guts structure in order to pass information like base
addresses, endianness etc between the init time and the runtime
operations (RCW override) which will get added in future patches.
There is no point in mapping and unmapping the DCFG CCSR space every
time we need to make a read, just map it once and keep its reference in
this new global struture.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v2->v3:
- fix error handling in fsl_guts_init() - iounmap() the CCSR range and
  set it to NULL on the "err" label rather than "err_nomem"
v1->v2: none
---
 drivers/soc/fsl/guts.c | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
index f87ee47c1503..a2b4c477b064 100644
--- a/drivers/soc/fsl/guts.c
+++ b/drivers/soc/fsl/guts.c
@@ -106,6 +106,11 @@ static const struct fsl_soc_die_attr fsl_soc_die[] = {
 	{ },
 };
 
+static struct fsl_soc_guts {
+	struct ccsr_guts __iomem *dcfg_ccsr;
+	bool little_endian;
+} soc;
+
 static const struct fsl_soc_die_attr *fsl_soc_die_match(
 	u32 svr, const struct fsl_soc_die_attr *matches)
 {
@@ -187,9 +192,7 @@ static int __init fsl_guts_init(void)
 	const struct fsl_soc_die_attr *soc_die;
 	const struct fsl_soc_data *soc_data;
 	const struct of_device_id *match;
-	struct ccsr_guts __iomem *regs;
 	struct device_node *np;
-	bool little_endian;
 	u64 soc_uid = 0;
 	u32 svr;
 	int ret;
@@ -199,24 +202,23 @@ static int __init fsl_guts_init(void)
 		return 0;
 	soc_data = match->data;
 
-	regs = of_iomap(np, DCFG_CCSR);
-	if (!regs) {
+	soc.dcfg_ccsr = of_iomap(np, DCFG_CCSR);
+	if (!soc.dcfg_ccsr) {
 		of_node_put(np);
 		return -ENOMEM;
 	}
 
-	little_endian = of_property_read_bool(np, "little-endian");
-	if (little_endian)
-		svr = ioread32(&regs->svr);
+	soc.little_endian = of_property_read_bool(np, "little-endian");
+	if (soc.little_endian)
+		svr = ioread32(&soc.dcfg_ccsr->svr);
 	else
-		svr = ioread32be(&regs->svr);
-	iounmap(regs);
+		svr = ioread32be(&soc.dcfg_ccsr->svr);
 	of_node_put(np);
 
 	/* Register soc device */
 	soc_dev_attr = kzalloc_obj(*soc_dev_attr);
 	if (!soc_dev_attr)
-		return -ENOMEM;
+		goto err_nomem;
 
 	ret = soc_attr_read_machine(soc_dev_attr);
 	if (ret)
@@ -269,6 +271,8 @@ static int __init fsl_guts_init(void)
 	kfree(soc_dev_attr->revision);
 	kfree(soc_dev_attr->serial_number);
 	kfree(soc_dev_attr);
+	iounmap(soc.dcfg_ccsr);
+	soc.dcfg_ccsr = NULL;
 
 	return ret;
 }
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 3/8] soc: fsl: guts: add a central fsl_guts_read() function

[Vladimir Oltean]

From: Ioana Ciornei <ioana.ciornei@nxp.com>

Add a central fsl_guts_read() function which will take into account the
endianness that was already determined. No point is duplicating the
if-else statement each time we need to read a DCFG register.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v1->v3: none
---
 drivers/soc/fsl/guts.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
index a2b4c477b064..e1b3a054e681 100644
--- a/drivers/soc/fsl/guts.c
+++ b/drivers/soc/fsl/guts.c
@@ -111,6 +111,14 @@ static struct fsl_soc_guts {
 	bool little_endian;
 } soc;
 
+static unsigned int fsl_guts_read(const void __iomem *reg)
+{
+	if (soc.little_endian)
+		return ioread32(reg);
+
+	return ioread32be(reg);
+}
+
 static const struct fsl_soc_die_attr *fsl_soc_die_match(
 	u32 svr, const struct fsl_soc_die_attr *matches)
 {
@@ -209,10 +217,7 @@ static int __init fsl_guts_init(void)
 	}
 
 	soc.little_endian = of_property_read_bool(np, "little-endian");
-	if (soc.little_endian)
-		svr = ioread32(&soc.dcfg_ccsr->svr);
-	else
-		svr = ioread32be(&soc.dcfg_ccsr->svr);
+	svr = fsl_guts_read(&soc.dcfg_ccsr->svr);
 	of_node_put(np);
 
 	/* Register soc device */
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 4/8] soc: fsl: guts: make it easier to determine on which SoC we are running

[Vladimir Oltean]

From: Ioana Ciornei <ioana.ciornei@nxp.com>

The guts driver will need to easily determine on which SoC it's running
when it will need to perform RCW override at runtime. The guts driver
knows this already because fsl_guts_init() reads the QorIQ/Layerscape
architectural System Version Register (SVR), but it doesn't save this
for later lookups.

Add a new qoriq_die enum to be used as an index in the fsl_soc_die
array. A new fsl_soc_die_match_one() function is also added so that we
can directly determine if the SVR is a match with a specific die.
The SVR value read from the DCFG CCSR is also kept in the global soc
structure so that it can be accessed when needed.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v1->v3: none
---
 drivers/soc/fsl/guts.c | 47 ++++++++++++++++++++++++++++++++++++------
 1 file changed, 41 insertions(+), 6 deletions(-)

diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
index e1b3a054e681..922560d98782 100644
--- a/drivers/soc/fsl/guts.c
+++ b/drivers/soc/fsl/guts.c
@@ -27,6 +27,23 @@ struct fsl_soc_data {
 	u32 uid_offset;
 };
 
+enum qoriq_die {
+	DIE_T4240,
+	DIE_T1040,
+	DIE_T2080,
+	DIE_T1024,
+	DIE_LS1043A,
+	DIE_LS2080A,
+	DIE_LS1088A,
+	DIE_LS1012A,
+	DIE_LS1046A,
+	DIE_LS2088A,
+	DIE_LS1021A,
+	DIE_LX2160A,
+	DIE_LS1028A,
+	DIE_MAX,
+};
+
 /* SoC die attribute definition for QorIQ platform */
 static const struct fsl_soc_die_attr fsl_soc_die[] = {
 	/*
@@ -34,21 +51,25 @@ static const struct fsl_soc_die_attr fsl_soc_die[] = {
 	 */
 
 	/* Die: T4240, SoC: T4240/T4160/T4080 */
+	[DIE_T4240] =
 	{ .die		= "T4240",
 	  .svr		= 0x82400000,
 	  .mask		= 0xfff00000,
 	},
 	/* Die: T1040, SoC: T1040/T1020/T1042/T1022 */
+	[DIE_T1040] =
 	{ .die		= "T1040",
 	  .svr		= 0x85200000,
 	  .mask		= 0xfff00000,
 	},
 	/* Die: T2080, SoC: T2080/T2081 */
+	[DIE_T2080] =
 	{ .die		= "T2080",
 	  .svr		= 0x85300000,
 	  .mask		= 0xfff00000,
 	},
 	/* Die: T1024, SoC: T1024/T1014/T1023/T1013 */
+	[DIE_T1024] =
 	{ .die		= "T1024",
 	  .svr		= 0x85400000,
 	  .mask		= 0xfff00000,
@@ -59,46 +80,55 @@ static const struct fsl_soc_die_attr fsl_soc_die[] = {
 	 */
 
 	/* Die: LS1043A, SoC: LS1043A/LS1023A */
+	[DIE_LS1043A] =
 	{ .die		= "LS1043A",
 	  .svr		= 0x87920000,
 	  .mask		= 0xffff0000,
 	},
 	/* Die: LS2080A, SoC: LS2080A/LS2040A/LS2085A */
+	[DIE_LS2080A] =
 	{ .die		= "LS2080A",
 	  .svr		= 0x87010000,
 	  .mask		= 0xff3f0000,
 	},
 	/* Die: LS1088A, SoC: LS1088A/LS1048A/LS1084A/LS1044A */
+	[DIE_LS1088A] =
 	{ .die		= "LS1088A",
 	  .svr		= 0x87030000,
 	  .mask		= 0xff3f0000,
 	},
 	/* Die: LS1012A, SoC: LS1012A */
+	[DIE_LS1012A] =
 	{ .die		= "LS1012A",
 	  .svr		= 0x87040000,
 	  .mask		= 0xffff0000,
 	},
 	/* Die: LS1046A, SoC: LS1046A/LS1026A */
+	[DIE_LS1046A] =
 	{ .die		= "LS1046A",
 	  .svr		= 0x87070000,
 	  .mask		= 0xffff0000,
 	},
 	/* Die: LS2088A, SoC: LS2088A/LS2048A/LS2084A/LS2044A */
+	[DIE_LS2088A] =
 	{ .die		= "LS2088A",
 	  .svr		= 0x87090000,
 	  .mask		= 0xff3f0000,
 	},
 	/* Die: LS1021A, SoC: LS1021A/LS1020A/LS1022A */
+	[DIE_LS1021A] =
 	{ .die		= "LS1021A",
 	  .svr		= 0x87000000,
 	  .mask		= 0xfff70000,
 	},
 	/* Die: LX2160A, SoC: LX2160A/LX2120A/LX2080A */
+	[DIE_LX2160A] =
 	{ .die          = "LX2160A",
 	  .svr          = 0x87360000,
 	  .mask         = 0xff3f0000,
 	},
 	/* Die: LS1028A, SoC: LS1028A */
+	[DIE_LS1028A] =
 	{ .die          = "LS1028A",
 	  .svr          = 0x870b0000,
 	  .mask         = 0xff3f0000,
@@ -109,6 +139,7 @@ static const struct fsl_soc_die_attr fsl_soc_die[] = {
 static struct fsl_soc_guts {
 	struct ccsr_guts __iomem *dcfg_ccsr;
 	bool little_endian;
+	u32 svr;
 } soc;
 
 static unsigned int fsl_guts_read(const void __iomem *reg)
@@ -119,11 +150,16 @@ static unsigned int fsl_guts_read(const void __iomem *reg)
 	return ioread32be(reg);
 }
 
+static bool fsl_soc_die_match_one(u32 svr, const struct fsl_soc_die_attr *match)
+{
+	return match->svr == (svr & match->mask);
+}
+
 static const struct fsl_soc_die_attr *fsl_soc_die_match(
 	u32 svr, const struct fsl_soc_die_attr *matches)
 {
 	while (matches->svr) {
-		if (matches->svr == (svr & matches->mask))
+		if (fsl_soc_die_match_one(svr, matches))
 			return matches;
 		matches++;
 	}
@@ -202,7 +238,6 @@ static int __init fsl_guts_init(void)
 	const struct of_device_id *match;
 	struct device_node *np;
 	u64 soc_uid = 0;
-	u32 svr;
 	int ret;
 
 	np = of_find_matching_node_and_match(NULL, fsl_guts_of_match, &match);
@@ -217,7 +252,7 @@ static int __init fsl_guts_init(void)
 	}
 
 	soc.little_endian = of_property_read_bool(np, "little-endian");
-	svr = fsl_guts_read(&soc.dcfg_ccsr->svr);
+	soc.svr = fsl_guts_read(&soc.dcfg_ccsr->svr);
 	of_node_put(np);
 
 	/* Register soc device */
@@ -229,7 +264,7 @@ static int __init fsl_guts_init(void)
 	if (ret)
 		of_machine_read_compatible(&soc_dev_attr->machine, 0);
 
-	soc_die = fsl_soc_die_match(svr, fsl_soc_die);
+	soc_die = fsl_soc_die_match(soc.svr, fsl_soc_die);
 	if (soc_die) {
 		soc_dev_attr->family = kasprintf(GFP_KERNEL, "QorIQ %s",
 						 soc_die->die);
@@ -239,12 +274,12 @@ static int __init fsl_guts_init(void)
 	if (!soc_dev_attr->family)
 		goto err_nomem;
 
-	soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "svr:0x%08x", svr);
+	soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "svr:0x%08x", soc.svr);
 	if (!soc_dev_attr->soc_id)
 		goto err_nomem;
 
 	soc_dev_attr->revision = kasprintf(GFP_KERNEL, "%d.%d",
-					   (svr >>  4) & 0xf, svr & 0xf);
+					   (soc.svr >>  4) & 0xf, soc.svr & 0xf);
 	if (!soc_dev_attr->revision)
 		goto err_nomem;
 
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 5/8] soc: fsl: guts: make fsl_soc_data available after fsl_guts_init()

[Vladimir Oltean]

In a future change, struct fsl_soc_data will be extended with methods
for performing RCW override.

Since this will be performed from a calling context outside
fsl_guts_init(), we need to keep track of the soc_data that we determine
at fsl_guts_init() time, so we can reference it later.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v2->v3: don't leave soc.data a valid pointer if fsl_guts_init() fails
v1->v2: none
---
 drivers/soc/fsl/guts.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
index 922560d98782..b9973ca3a440 100644
--- a/drivers/soc/fsl/guts.c
+++ b/drivers/soc/fsl/guts.c
@@ -138,6 +138,7 @@ static const struct fsl_soc_die_attr fsl_soc_die[] = {
 
 static struct fsl_soc_guts {
 	struct ccsr_guts __iomem *dcfg_ccsr;
+	const struct fsl_soc_data *data;
 	bool little_endian;
 	u32 svr;
 } soc;
@@ -231,10 +232,9 @@ static const struct of_device_id fsl_guts_of_match[] = {
 
 static int __init fsl_guts_init(void)
 {
-	struct soc_device_attribute *soc_dev_attr;
+	struct soc_device_attribute *soc_dev_attr = NULL;
 	static struct soc_device *soc_dev;
 	const struct fsl_soc_die_attr *soc_die;
-	const struct fsl_soc_data *soc_data;
 	const struct of_device_id *match;
 	struct device_node *np;
 	u64 soc_uid = 0;
@@ -243,12 +243,12 @@ static int __init fsl_guts_init(void)
 	np = of_find_matching_node_and_match(NULL, fsl_guts_of_match, &match);
 	if (!np)
 		return 0;
-	soc_data = match->data;
+	soc.data = match->data;
 
 	soc.dcfg_ccsr = of_iomap(np, DCFG_CCSR);
 	if (!soc.dcfg_ccsr) {
 		of_node_put(np);
-		return -ENOMEM;
+		goto err_nomem;
 	}
 
 	soc.little_endian = of_property_read_bool(np, "little-endian");
@@ -283,9 +283,9 @@ static int __init fsl_guts_init(void)
 	if (!soc_dev_attr->revision)
 		goto err_nomem;
 
-	if (soc_data)
-		soc_uid = fsl_guts_get_soc_uid(soc_data->sfp_compat,
-					       soc_data->uid_offset);
+	if (soc.data)
+		soc_uid = fsl_guts_get_soc_uid(soc.data->sfp_compat,
+					       soc.data->uid_offset);
 	if (soc_uid)
 		soc_dev_attr->serial_number = kasprintf(GFP_KERNEL, "%016llX",
 							soc_uid);
@@ -311,8 +311,11 @@ static int __init fsl_guts_init(void)
 	kfree(soc_dev_attr->revision);
 	kfree(soc_dev_attr->serial_number);
 	kfree(soc_dev_attr);
-	iounmap(soc.dcfg_ccsr);
-	soc.dcfg_ccsr = NULL;
+	if (soc.dcfg_ccsr) {
+		iounmap(soc.dcfg_ccsr);
+		soc.dcfg_ccsr = NULL;
+	}
+	soc.data = NULL;
 
 	return ret;
 }
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 6/8] dt-bindings: fsl: layerscape-dcfg: define DCFG_DCSR region

[Vladimir Oltean]

In Layerscape (Arm) and QorIQ (PowerPC) devices, hardware peripherals
are accessed by the CPU through a portion of the SoC address space
called CCSR ("Configuration, Control, and Status Registers"). All
hardware IP blocks have their registers mapped here, and the Device
Configuration block makes no exception.

However, there exists a secondary range of the address space named DCSR
("Debug Control and Status Registers") which, like CCSR, also holds
registers of hardware IP blocks, except the DCSR contents is hidden in
all public reference manuals.

The intention of the CCSR/DCSR split, to the best of my knowledge, was
to place the functionality that is too low level for normal use, and
which is necessary only for debug, in a completely separate address
space which can be hidden.

A use case has appeared where networking SerDes lanes need to be
reconfigured at runtime for a different protocol (example: 10GBase-R to
SGMII), and the architecture of the SoCs does not normally permit that.
The Reset Configuration Word (RCW) is a data structure read by the SoC
preboot loader (PBL) which contains stuff like pinmuxing and SerDes
protocol mapping for each lane.

The RCW that the PBL has loaded is visible in the DCFG block's normal
status registers (from CCSR), as read only. Turns out, the RCW is also
mapped in the DCFG's shadow register map (in DCSR), in a write-only
form. Writing to the RCW registers from the DCFG's DCSR space to change
what the PBL has loaded is called "RCW override".

It has been validated that the RCW override procedure is necessary to
reconfigure the networking data path when a SerDes lane performs a major
protocol change. It changes some internal muxes which connect the PCS to
either the 10G MAC or to the 1G MAC.

Defining the DCSR area of the DCFG as a secondary 'reg' array element
allows operating systems to perform RCW overrides. Since it is
introduced late in the binding's lifetime, it is optional. It can be
identified by name, but also by index (first 'reg' is CCSR).

Note that while all SoCs should have a DCFG register block in DCSR, we
only need to expose it for the SoCs where the RCW override procedure is
known to be needed and has been validated.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
---
Cc: Conor Dooley <conor+dt@kernel.org>
Cc: Krzysztof Kozlowski <krzk+dt@kernel.org>
Cc: Rob Herring <robh@kernel.org>

v2->v3: none
v1->v2:
- add Conor's review tag
- update email addresses of DT maintainers
---
 .../bindings/soc/fsl/fsl,layerscape-dcfg.yaml     | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml b/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml
index 3fb0534ea597..fc14fd0bf84b 100644
--- a/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml
+++ b/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml
@@ -36,7 +36,20 @@ properties:
           - const: simple-mfd
 
   reg:
-    maxItems: 1
+    minItems: 1
+    items:
+      - description:
+          Customer-visible DCFG register map from CCSR address space
+          (Configuration, Control and Status Registers)
+      - description:
+          Customer-hidden DCFG register map from DCSR address space
+          (Debug Control and Status Registers)
+
+  reg-names:
+    minItems: 1
+    items:
+      - const: dcfg_ccsr
+      - const: dcfg_dcsr
 
   little-endian: true
   big-endian: true
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 7/8] soc: fsl: guts: implement the RCW override procedure

[Vladimir Oltean]

From: Ioana Ciornei <ioana.ciornei@nxp.com>

Add support for the RCW override procedure which enables runtime
reconfiguration of the protocol running on a SerDes lane. The procedure
is done through the DCFG DCSR space which now can be defined as the
second memory region of the guts DT node.
Support is added on the following SoCs: LS1046A, LS1088A, LS2088A.

The procedure is exported to the "client" driver - the Lynx10G SerDes
PHY driver - through the following functions:
- fsl_guts_lane_init() used to notify the initial / boot time lane mode
  running on a SerDes lane.
- fsl_guts_lane_validate() used to validate that changing the protocol
  on a specific lane is supported.
- fsl_guts_lane_set_mode() which can be used to request the RCW
  procedure be executed for a specific lane.

Since the RCW override procedure is different depending on the SoC, the
private fsl_soc_data structure is updated with two new per SoC callbacks
(.serdes_get_rcw_override() and .serdes_init_rcwcr()) which get used
from the generic fsl_guts_lane_set_mode() function. These two callbacks
hide all the SoC specific register offsets, masks and values so that the
_set_mode() procedure is straightforward.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v2->v3:
- move iounmap() of dcfg_dcsr area on the 'err' label
- use __ffs() (standard API which also works on armv7) instead of
  __bf_shf() (an internal helper meant for compile-time constants in
  bitfield.h)
- only operate on lanes on which fsl_guts_lane_init() was called in
  ls2088a_serdes_init_rcwcr()
- use read_poll_timeout_atomic() in fsl_guts_rcw_rmw() to clarify that
  DCFG_DCSR writes are supposed to reflect back in DCFG_CCSR
- check soc.data in fsl_guts_serdes_get_rcw_override() to ensure
  fsl_guts_init() succeeded
- put parentheses around (lane) in LS1088A_RCWSR29_SRDS_PRTCL_S1_LNn()
  and LS1088A_RCWSR30_SRDS_PRTCL_S2_LNn() macro expressions

v1->v2:
- drop DT maintainers from explicit CC
- keep devicetree@vger.kernel.org CCed on entire series
- include missing <linux/bitfield.h>
- namespace SRDS_PRTCL values for LS1046A and LS1088A, even if they are
  the same. For LS1028A (not covered here) they are not.
- prefix SRDS_CLK_SEL_{GMII,XGMII} with LS2088A_
- reorder alphanumerically (LS1046A should come before LS1088A)
---
 drivers/soc/fsl/guts.c   | 343 ++++++++++++++++++++++++++++++++++++++-
 include/linux/fsl/guts.h |  20 ++-
 2 files changed, 357 insertions(+), 6 deletions(-)

diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
index b9973ca3a440..4d49d1b95300 100644
--- a/drivers/soc/fsl/guts.c
+++ b/drivers/soc/fsl/guts.c
@@ -5,7 +5,10 @@
  * Copyright (C) 2016 Freescale Semiconductor, Inc.
  */
 
+#include <linux/bitfield.h>
+#include <linux/bitops.h>
 #include <linux/io.h>
+#include <linux/iopoll.h>
 #include <linux/slab.h>
 #include <linux/module.h>
 #include <linux/of_fdt.h>
@@ -15,6 +18,30 @@
 #include <linux/fsl/guts.h>
 
 #define DCFG_CCSR	0
+#define DCFG_DCSR	1
+
+#define MAX_NUM_LANES	8
+#define MAX_NUM_SERDES	2
+
+#define RCW_TIMEOUT_US	1
+
+#define LS1046A_RCWSR5_SRDS_PRTCL_S1(lane)	\
+	GENMASK(19 + 4 * (lane), 16 + 4 * (lane))
+#define LS1046A_SRDS_PRTCL_XFI				1
+#define LS1046A_SRDS_PRTCL_100BASEX_SGMII		3
+
+#define LS1088A_RCWSR29_SRDS_PRTCL_S1_LNn(lane)	\
+	GENMASK(19 + 4 * (3 - (lane)), 16 + 4 * (3 - (lane)))
+#define LS1088A_RCWSR30_SRDS_PRTCL_S2_LNn(lane)	\
+	GENMASK(3 + 4 * (3 - (lane)), 4 * (3 - (lane)))
+#define LS1088A_SRDS_PRTCL_XFI				1
+#define LS1088A_SRDS_PRTCL_100BASEX_SGMII		3
+
+#define LS2088A_RCWSR30_SRDS_CLK_EN_SEL_XGMII_S1	BIT(14)
+#define LS2088A_RCWSR30_SRDS_CLK_SEL_XGMII_Ln_S1(lane)	BIT(6 + (7 - (lane)))
+#define LS2088A_RCWSR30_SRDS_CLK_SEL_MSK		GENMASK(13, 6)
+#define LS2088A_SRDS_CLK_SEL_XGMII			1
+#define LS2088A_SRDS_CLK_SEL_GMII			0
 
 struct fsl_soc_die_attr {
 	char	*die;
@@ -22,9 +49,20 @@ struct fsl_soc_die_attr {
 	u32	mask;
 };
 
+struct fsl_soc_serdes_rcw_override {
+	int offset;
+	int mask;
+	int val;
+};
+
 struct fsl_soc_data {
 	const char *sfp_compat;
 	u32 uid_offset;
+	int num_serdes;
+	int (*serdes_get_rcw_override)(int index, int lane,
+				       enum lynx_lane_mode lane_mode,
+				       struct fsl_soc_serdes_rcw_override *override);
+	int (*serdes_init_rcwcr)(int index);
 };
 
 enum qoriq_die {
@@ -138,9 +176,14 @@ static const struct fsl_soc_die_attr fsl_soc_die[] = {
 
 static struct fsl_soc_guts {
 	struct ccsr_guts __iomem *dcfg_ccsr;
+	struct ccsr_guts __iomem *dcfg_dcsr;
 	const struct fsl_soc_data *data;
 	bool little_endian;
 	u32 svr;
+	enum lynx_lane_mode lane_mode[MAX_NUM_SERDES][MAX_NUM_LANES];
+	unsigned long lanes_initialized[MAX_NUM_SERDES];
+	bool rcwcr_init_done;
+	spinlock_t rcwcr_lock; /* serializes concurrent writes to the RCWCR */
 } soc;
 
 static unsigned int fsl_guts_read(const void __iomem *reg)
@@ -151,6 +194,33 @@ static unsigned int fsl_guts_read(const void __iomem *reg)
 	return ioread32be(reg);
 }
 
+static void fsl_guts_write(void __iomem *reg, u32 val)
+{
+	if (soc.little_endian)
+		iowrite32(val, reg);
+	else
+		iowrite32be(val, reg);
+}
+
+/* Some fields of the Reset Configuration Word (RCW) can be overridden at
+ * runtime by writing to the RCWCRn registers contained within the DCSR space
+ * of the Device Configuration (DCFG) block. The layout of the RCWCRn registers
+ * is identical with the read-only RCWSRn from the CCSR space.
+ */
+static int fsl_guts_rcw_rmw(int offset, u32 val, u32 mask)
+{
+	u32 rcwcr, rcwsr = fsl_guts_read(&soc.dcfg_ccsr->rcwsr[offset]);
+
+	rcwcr = rcwsr & ~mask;
+	rcwcr |= val;
+	fsl_guts_write(&soc.dcfg_dcsr->rcwcr[offset], rcwcr);
+
+	/* Updates to RCWCR should be visible back in RCWSR */
+	return read_poll_timeout_atomic(fsl_guts_read, rcwsr, rcwsr == rcwcr,
+					0, RCW_TIMEOUT_US, false,
+					&soc.dcfg_ccsr->rcwsr[offset]);
+}
+
 static bool fsl_soc_die_match_one(u32 svr, const struct fsl_soc_die_attr *match)
 {
 	return match->svr == (svr & match->mask);
@@ -167,6 +237,131 @@ static const struct fsl_soc_die_attr *fsl_soc_die_match(
 	return NULL;
 }
 
+static int
+fsl_guts_serdes_get_rcw_override(int serdes_idx, int lane,
+				 enum lynx_lane_mode lane_mode,
+				 struct fsl_soc_serdes_rcw_override *override)
+{
+	const struct fsl_soc_data *soc_data = soc.data;
+
+	if (!soc_data)
+		return -ENODEV;
+
+	if (serdes_idx >= soc_data->num_serdes || serdes_idx <= 0)
+		return -ERANGE;
+
+	if (lane >= MAX_NUM_LANES || lane < 0)
+		return -ERANGE;
+
+	if ((!fsl_soc_die_match_one(soc.svr, &fsl_soc_die[DIE_LS1088A]) &&
+	     !fsl_soc_die_match_one(soc.svr, &fsl_soc_die[DIE_LS2088A]) &&
+	     !fsl_soc_die_match_one(soc.svr, &fsl_soc_die[DIE_LS1046A])) ||
+	    !soc_data->serdes_get_rcw_override) {
+		pr_debug("RCW override not implemented for SoC\n");
+		return -EINVAL;
+	}
+
+	if (!soc.dcfg_dcsr) {
+		pr_debug("Device tree does not define DCFG_DCSR region necessary for RCW override\n");
+		return -EINVAL;
+	}
+
+	return soc_data->serdes_get_rcw_override(serdes_idx, lane, lane_mode,
+						 override);
+}
+
+/**
+ * fsl_guts_lane_validate() - Validate that SerDes protocol is implemented and
+ *	supported on current SoC
+ * @serdes_idx: one-based SerDes block index
+ * @lane: zero-based lane index within SerDes
+ * @lane_mode: requested SerDes protocol
+ *
+ * Should be called before actually requesting the RCW override procedure to be
+ * applied using %fsl_guts_lane_set_mode()
+ *
+ * Return: 0 if RCW override to protocol is possible, negative error otherwise
+ */
+int fsl_guts_lane_validate(int serdes_idx, int lane, enum lynx_lane_mode lane_mode)
+{
+	struct fsl_soc_serdes_rcw_override override;
+
+	return fsl_guts_serdes_get_rcw_override(serdes_idx, lane, lane_mode,
+						&override);
+}
+EXPORT_SYMBOL_NS_GPL(fsl_guts_lane_validate, "FSL_GUTS");
+
+/**
+ * fsl_guts_lane_init() - Notify guts module of current SerDes lane configuration
+ * @serdes_idx: one-based SerDes block index
+ * @lane: zero-based lane index within SerDes
+ * @lane_mode: initial / boot time SerDes protocol for lane
+ *
+ * On the LS208xA SoC, the RCW override procedure needs to be aware of all link
+ * modes which are configured on a SerDes block.
+ *
+ * Return: 0 if passed parameters are known to driver, negative error otherwise
+ */
+int fsl_guts_lane_init(int serdes_idx, int lane, enum lynx_lane_mode lane_mode)
+{
+	int err;
+
+	err = fsl_guts_lane_validate(serdes_idx, lane, lane_mode);
+	if (err)
+		return err;
+
+	soc.lane_mode[serdes_idx - 1][lane] = lane_mode;
+	soc.lanes_initialized[serdes_idx - 1] |= BIT(lane);
+
+	return 0;
+}
+EXPORT_SYMBOL_NS_GPL(fsl_guts_lane_init, "FSL_GUTS");
+
+/**
+ * fsl_guts_lane_set_mode() - apply RCW override procedure for SerDes lane
+ * @serdes_idx: one-based SerDes block index
+ * @lane: zero-based lane index within SerDes
+ * @lane_mode: requested SerDes protocol
+ *
+ * Return: 0 on success, negative error otherwise
+ */
+int fsl_guts_lane_set_mode(int serdes_idx, int lane, enum lynx_lane_mode lane_mode)
+{
+	struct fsl_soc_serdes_rcw_override override;
+	int err;
+
+	err = fsl_guts_serdes_get_rcw_override(serdes_idx, lane, lane_mode,
+					       &override);
+	if (err)
+		return err;
+
+	/* Ensure fsl_guts_lane_init() was previously called */
+	if (!(soc.lanes_initialized[serdes_idx - 1] & BIT(lane)))
+		return -EINVAL;
+
+	spin_lock(&soc.rcwcr_lock);
+
+	if (soc.data->serdes_init_rcwcr) {
+		err = soc.data->serdes_init_rcwcr(serdes_idx);
+		if (err)
+			goto out_unlock;
+	}
+
+	err = fsl_guts_rcw_rmw(override.offset,
+			       override.val << __ffs(override.mask),
+			       override.mask);
+	if (err)
+		pr_err("RCW override failed: %pe\n", ERR_PTR(err));
+	else
+		soc.lane_mode[serdes_idx - 1][lane] = lane_mode;
+
+out_unlock:
+	spin_unlock(&soc.rcwcr_lock);
+
+	return err;
+}
+EXPORT_SYMBOL_NS_GPL(fsl_guts_lane_set_mode, "FSL_GUTS");
+
 static u64 fsl_guts_get_soc_uid(const char *compat, unsigned int offset)
 {
 	struct device_node *np;
@@ -193,9 +388,143 @@ static u64 fsl_guts_get_soc_uid(const char *compat, unsigned int offset)
 	return uid;
 }
 
+static int ls1046a_serdes_get_rcw_override(int index, int lane,
+					   enum lynx_lane_mode lane_mode,
+					   struct fsl_soc_serdes_rcw_override *override)
+{
+	/* The RCW override procedure has to write to different registers
+	 * depending on the SerDes block index.
+	 */
+	switch (index) {
+	case 1:
+		override->offset = 4;
+		override->mask = LS1046A_RCWSR5_SRDS_PRTCL_S1(lane);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (lynx_lane_mode_uses_xgmii_mac(lane_mode))
+		override->val = LS1046A_SRDS_PRTCL_XFI;
+	else if (lynx_lane_mode_uses_gmii_mac(lane_mode))
+		override->val = LS1046A_SRDS_PRTCL_100BASEX_SGMII;
+	else
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ls1088a_serdes_get_rcw_override(int index, int lane,
+					   enum lynx_lane_mode lane_mode,
+					   struct fsl_soc_serdes_rcw_override *override)
+{
+	/* The RCW override procedure has to write to different registers
+	 * depending on the SerDes block index.
+	 */
+	switch (index) {
+	case 1:
+		override->offset = 28;
+		override->mask = LS1088A_RCWSR29_SRDS_PRTCL_S1_LNn(lane);
+		break;
+	case 2:
+		override->offset = 29;
+		override->mask = LS1088A_RCWSR30_SRDS_PRTCL_S2_LNn(lane);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (lynx_lane_mode_uses_xgmii_mac(lane_mode))
+		override->val = LS1088A_SRDS_PRTCL_XFI;
+	else if (lynx_lane_mode_uses_gmii_mac(lane_mode))
+		override->val = LS1088A_SRDS_PRTCL_100BASEX_SGMII;
+	else
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ls2088a_serdes_get_rcw_override(int index, int lane,
+					   enum lynx_lane_mode lane_mode,
+					   struct fsl_soc_serdes_rcw_override *override)
+{
+	switch (index) {
+	case 1:
+		override->offset = 29;
+		override->mask = LS2088A_RCWSR30_SRDS_CLK_SEL_XGMII_Ln_S1(lane);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (lynx_lane_mode_uses_xgmii_mac(lane_mode))
+		override->val = LS2088A_SRDS_CLK_SEL_XGMII;
+	else if (lynx_lane_mode_uses_gmii_mac(lane_mode))
+		override->val = LS2088A_SRDS_CLK_SEL_GMII;
+	else
+		return -EINVAL;
+
+	return 0;
+}
+
+static int ls2088a_serdes_init_rcwcr(int serdes_idx)
+{
+	int i, err;
+	u32 reg;
+
+	/* SerDes 2 supports only SGMII for networking. There should be
+	 * no need for RCW override
+	 */
+	if (serdes_idx != 1)
+		return -EINVAL;
+
+	if (soc.rcwcr_init_done)
+		return 0;
+
+	/* SRDS_CLK_EN_SEL_XGMII_S1: SerDes Clock Enable Select XGMII Serdes 1:
+	 * Enables to select GMII/XGMII clock according to
+	 * SRDS_CLK_SEL_XGMII_Ln_S1
+	 */
+	reg = LS2088A_RCWSR30_SRDS_CLK_EN_SEL_XGMII_S1;
+
+	/* We need to configure the initial state of all lanes for
+	 * the SerDes block #1
+	 */
+	for_each_set_bit(i, &soc.lanes_initialized[serdes_idx - 1], MAX_NUM_LANES)
+		if (lynx_lane_mode_uses_xgmii_mac(soc.lane_mode[serdes_idx - 1][i]))
+			reg |= LS2088A_RCWSR30_SRDS_CLK_SEL_XGMII_Ln_S1(i);
+
+	err = fsl_guts_rcw_rmw(29, reg,
+			       LS2088A_RCWSR30_SRDS_CLK_EN_SEL_XGMII_S1 |
+			       LS2088A_RCWSR30_SRDS_CLK_SEL_MSK);
+	if (err)
+		return err;
+
+	soc.rcwcr_init_done = true;
+
+	return 0;
+}
+
+static const struct fsl_soc_data ls1088a_data = {
+	.serdes_get_rcw_override = ls1088a_serdes_get_rcw_override,
+	.num_serdes = 2,
+};
+
+static const struct fsl_soc_data ls1046a_data = {
+	.serdes_get_rcw_override = ls1046a_serdes_get_rcw_override,
+	.num_serdes = 2,
+};
+
+static const struct fsl_soc_data ls2088a_data = {
+	.serdes_get_rcw_override = ls2088a_serdes_get_rcw_override,
+	.serdes_init_rcwcr = ls2088a_serdes_init_rcwcr,
+	.num_serdes = 2,
+};
+
 static const struct fsl_soc_data ls1028a_data = {
 	.sfp_compat = "fsl,ls1028a-sfp",
 	.uid_offset = 0x21c,
+	.num_serdes = 1,
 };
 
 /*
@@ -221,10 +550,10 @@ static const struct of_device_id fsl_guts_of_match[] = {
 	{ .compatible = "fsl,mpc8572-guts", },
 	{ .compatible = "fsl,ls1021a-dcfg", },
 	{ .compatible = "fsl,ls1043a-dcfg", },
-	{ .compatible = "fsl,ls2080a-dcfg", },
-	{ .compatible = "fsl,ls1088a-dcfg", },
+	{ .compatible = "fsl,ls2080a-dcfg", .data = &ls2088a_data},
+	{ .compatible = "fsl,ls1088a-dcfg", .data = &ls1088a_data},
 	{ .compatible = "fsl,ls1012a-dcfg", },
-	{ .compatible = "fsl,ls1046a-dcfg", },
+	{ .compatible = "fsl,ls1046a-dcfg", .data = &ls1046a_data},
 	{ .compatible = "fsl,lx2160a-dcfg", },
 	{ .compatible = "fsl,ls1028a-dcfg", .data = &ls1028a_data},
 	{}
@@ -250,6 +579,8 @@ static int __init fsl_guts_init(void)
 		of_node_put(np);
 		goto err_nomem;
 	}
+	/* DCFG_DCSR is optional */
+	soc.dcfg_dcsr = of_iomap(np, DCFG_DCSR);
 
 	soc.little_endian = of_property_read_bool(np, "little-endian");
 	soc.svr = fsl_guts_read(&soc.dcfg_ccsr->svr);
@@ -296,6 +627,8 @@ static int __init fsl_guts_init(void)
 		goto err;
 	}
 
+	spin_lock_init(&soc.rcwcr_lock);
+
 	pr_info("Machine: %s\n", soc_dev_attr->machine);
 	pr_info("SoC family: %s\n", soc_dev_attr->family);
 	pr_info("SoC ID: %s, Revision: %s\n",
@@ -311,6 +644,10 @@ static int __init fsl_guts_init(void)
 	kfree(soc_dev_attr->revision);
 	kfree(soc_dev_attr->serial_number);
 	kfree(soc_dev_attr);
+	if (soc.dcfg_dcsr) {
+		iounmap(soc.dcfg_dcsr);
+		soc.dcfg_dcsr = NULL;
+	}
 	if (soc.dcfg_ccsr) {
 		iounmap(soc.dcfg_ccsr);
 		soc.dcfg_ccsr = NULL;
diff --git a/include/linux/fsl/guts.h b/include/linux/fsl/guts.h
index fdb55ca47a4f..605300908a35 100644
--- a/include/linux/fsl/guts.h
+++ b/include/linux/fsl/guts.h
@@ -13,6 +13,7 @@
 
 #include <linux/types.h>
 #include <linux/io.h>
+#include <soc/fsl/phy-fsl-lynx.h>
 
 /*
  * Global Utility Registers.
@@ -91,9 +92,15 @@ struct ccsr_guts {
 	u32	iovselsr;	/* 0x.00c0 - I/O voltage select status register
 					     Called 'elbcvselcr' on 86xx SOCs */
 	u8	res0c4[0x100 - 0xc4];
-	u32	rcwsr[16];	/* 0x.0100 - Reset Control Word Status registers
-					     There are 16 registers */
-	u8	res140[0x224 - 0x140];
+	/* 0x.0100 - read-only Reset Configuration Word Status registers in
+	 * CCSR, or write-only Reset Configuration Word Control registers in
+	 * DCSR. In both cases there are 32 registers.
+	 */
+	union {
+		u32	rcwsr[32];
+		u32	rcwcr[32];
+	};
+	u8	res180[0x224 - 0x180];
 	u32	iodelay1;	/* 0x.0224 - IO delay control register 1 */
 	u32	iodelay2;	/* 0x.0228 - IO delay control register 2 */
 	u8	res22c[0x604 - 0x22c];
@@ -131,6 +138,13 @@ struct ccsr_guts {
 	u32	srds2cr1;	/* 0x.0f44 - SerDes2 Control Register 0 */
 } __attribute__ ((packed));
 
+int fsl_guts_lane_init(int serdes_idx, int lane,
+		       enum lynx_lane_mode lane_mode);
+int fsl_guts_lane_validate(int serdes_idx, int lane,
+			   enum lynx_lane_mode lane_mode);
+int fsl_guts_lane_set_mode(int serdes_idx, int lane,
+			   enum lynx_lane_mode lane_mode);
+
 /* Alternate function signal multiplex control */
 #define MPC85xx_PMUXCR_QE(x) (0x8000 >> (x))
 
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

[PATCH v3 phy-next 8/8] phy: lynx-10g: use RCW override procedure for dynamic protocol change

[Vladimir Oltean]

Up until this patch, the only protocol change supported was between
1000Base-X/SGMII and 2500Base-X. The others require an RCW override
procedure which was lacking.

Since now the guts driver provides the means of applying this procedure,
make use of it and remove any comment which mentioned the limitation.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
v1->v3: none
---
 drivers/phy/freescale/Kconfig            |  1 +
 drivers/phy/freescale/phy-fsl-lynx-10g.c | 24 +++++++++++++++---------
 2 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/drivers/phy/freescale/Kconfig b/drivers/phy/freescale/Kconfig
index 5bf3864fbe64..d4e189fffbf8 100644
--- a/drivers/phy/freescale/Kconfig
+++ b/drivers/phy/freescale/Kconfig
@@ -58,6 +58,7 @@ config PHY_FSL_LYNX_10G
 	tristate "Freescale Layerscape Lynx 10G SerDes PHY support"
 	depends on OF
 	depends on ARCH_LAYERSCAPE || COMPILE_TEST
+	select FSL_GUTS
 	select GENERIC_PHY
 	select PHY_FSL_LYNX_CORE
 	help
diff --git a/drivers/phy/freescale/phy-fsl-lynx-10g.c b/drivers/phy/freescale/phy-fsl-lynx-10g.c
index 38def160ef1a..5ece7889aed7 100644
--- a/drivers/phy/freescale/phy-fsl-lynx-10g.c
+++ b/drivers/phy/freescale/phy-fsl-lynx-10g.c
@@ -8,6 +8,7 @@
 #include <linux/phy/phy.h>
 #include <linux/platform_device.h>
 #include <linux/workqueue.h>
+#include <linux/fsl/guts.h>
 
 #include "phy-fsl-lynx-core.h"
 
@@ -446,6 +447,7 @@ static void lynx_10g_lane_read_configuration(struct lynx_lane *lane)
 	}
 
 	lynx_10g_backup_pccr_val(lane);
+	fsl_guts_lane_init(priv->info->index, lane->id, lane->mode);
 }
 
 static int ls1028a_get_pccr(enum lynx_lane_mode lane_mode, int lane,
@@ -1167,14 +1169,7 @@ static bool lynx_10g_lane_mode_needs_rcw_override(struct lynx_lane *lane,
 
 	/* Major protocol changes, which involve changing the PCS connection to
 	 * the GMII MAC with the one to the XGMII MAC, require an RCW override
-	 * procedure to reconfigure an internal mux, as documented here:
-	 * https://lore.kernel.org/linux-phy/20230810102631.bvozjer3t67r67iy@skbuf/
-	 * This is SoC-specific, and not yet implemented in drivers/soc/fsl/guts.c.
-	 *
-	 * So the supported set of protocols depends on the initial lane mode.
-	 *
-	 * Minor protocol changes (SGMII <-> 1000Base-X <-> 2500Base-X or
-	 * 10GBase-R <-> USXGMII) are supported.
+	 * procedure to reconfigure an internal mux.
 	 */
 	if ((lynx_lane_mode_uses_gmii_mac(curr) &&
 	     lynx_lane_mode_uses_xgmii_mac(new)) ||
@@ -1189,6 +1184,7 @@ static int lynx_10g_validate(struct phy *phy, enum phy_mode mode, int submode,
 			     union phy_configure_opts *opts)
 {
 	struct lynx_lane *lane = phy_get_drvdata(phy);
+	struct lynx_priv *priv = lane->priv;
 	enum lynx_lane_mode lane_mode;
 	int err;
 
@@ -1197,7 +1193,8 @@ static int lynx_10g_validate(struct phy *phy, enum phy_mode mode, int submode,
 		return err;
 
 	if (lynx_10g_lane_mode_needs_rcw_override(lane, lane_mode))
-		return -EINVAL;
+		return fsl_guts_lane_validate(priv->info->index, lane->id,
+					      lane_mode);
 
 	return 0;
 }
@@ -1205,6 +1202,7 @@ static int lynx_10g_validate(struct phy *phy, enum phy_mode mode, int submode,
 static int lynx_10g_set_mode(struct phy *phy, enum phy_mode mode, int submode)
 {
 	struct lynx_lane *lane = phy_get_drvdata(phy);
+	struct lynx_priv *priv = lane->priv;
 	bool powered_up = lane->powered_up;
 	enum lynx_lane_mode lane_mode;
 	int err;
@@ -1225,6 +1223,13 @@ static int lynx_10g_set_mode(struct phy *phy, enum phy_mode mode, int submode)
 	if (powered_up)
 		lynx_10g_lane_halt(phy);
 
+	if (lynx_10g_lane_mode_needs_rcw_override(lane, lane_mode)) {
+		err = fsl_guts_lane_set_mode(priv->info->index, lane->id,
+					     lane_mode);
+		if (err)
+			goto out;
+	}
+
 	err = lynx_10g_lane_disable_pcvt(lane, lane->mode);
 	if (err)
 		goto out;
@@ -1314,6 +1319,7 @@ static struct platform_driver lynx_10g_driver = {
 };
 module_platform_driver(lynx_10g_driver);
 
+MODULE_IMPORT_NS("FSL_GUTS");
 MODULE_IMPORT_NS("PHY_FSL_LYNX");
 MODULE_AUTHOR("Ioana Ciornei <ioana.ciornei@nxp.com>");
 MODULE_AUTHOR("Vladimir Oltean <vladimir.oltean@nxp.com>");
-- 
2.34.1


-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Dereferencing a NULL pointer during error cleanup if the initial allocation of `soc_dev_attr` fails.
--

--- Patch [2]: [PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state ---
commit dce21fc9192bcf2ec7c5c2e9639e43b7a6c53c57
Author: Ioana Ciornei <ioana.ciornei@nxp.com>

soc: fsl: guts: add a global structure to hold state

This commit introduces the fsl_soc_guts global structure to hold the DCFG CCSR
mapping and endianness state. It maps the memory once at initialization so it
can be reused during runtime operations instead of repeatedly mapping and
unmapping the space.

> diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
> index f87ee47c15035..a2b4c477b064b 100644
> --- a/drivers/soc/fsl/guts.c
> +++ b/drivers/soc/fsl/guts.c
[ ... ]
> @@ -199,24 +202,23 @@ static int __init fsl_guts_init(void)
[ ... ]
>  	/* Register soc device */
>  	soc_dev_attr = kzalloc_obj(*soc_dev_attr);
>  	if (!soc_dev_attr)
> -		return -ENOMEM;
> +		goto err_nomem;

[Severity: High]
Will jumping to err_nomem cause a NULL pointer dereference?

If the allocation of soc_dev_attr fails, the code now jumps to the error
handling path which dereferences soc_dev_attr to free its members:

> @@ -269,6 +271,8 @@ static int __init fsl_guts_init(void)
>  	kfree(soc_dev_attr->revision);
>  	kfree(soc_dev_attr->serial_number);

Since soc_dev_attr is NULL when this allocation fails, won't evaluating
soc_dev_attr->revision and other members result in a kernel panic?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=2

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 3/8] soc: fsl: guts: add a central fsl_guts_read() function

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Dereferencing NULL pointer 'soc_dev_attr' on the error path if its memory allocation fails.
--

--- Patch [3]: [PATCH v3 phy-next 3/8] soc: fsl: guts: add a central fsl_guts_read() function ---
commit cc1312c02ab5c2cd3fed5e2907608e44453c7816
Author: Ioana Ciornei <ioana.ciornei@nxp.com>

soc: fsl: guts: add a central fsl_guts_read() function

This commit adds a central fsl_guts_read() function that handles reading
DCFG registers based on the already determined endianness. This avoids
duplicating the if-else endianness checks across the driver.

> diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
> index a2b4c477b064b..e1b3a054e681b 100644
> --- a/drivers/soc/fsl/guts.c
> +++ b/drivers/soc/fsl/guts.c
> @@ -111,6 +111,14 @@ static struct fsl_soc_guts {
>  	bool little_endian;
>  } soc;
>  
> +static unsigned int fsl_guts_read(const void __iomem *reg)
> +{
> +	if (soc.little_endian)
> +		return ioread32(reg);
> +
> +	return ioread32be(reg);
> +}
> +
>  static const struct fsl_soc_die_attr *fsl_soc_die_match(
>  	u32 svr, const struct fsl_soc_die_attr *matches)
>  {

[ ... ]

> @@ -209,10 +217,7 @@ static int __init fsl_guts_init(void)
>  	}
>  
>  	soc.little_endian = of_property_read_bool(np, "little-endian");
> -	if (soc.little_endian)
> -		svr = ioread32(&soc.dcfg_ccsr->svr);
> -	else
> -		svr = ioread32be(&soc.dcfg_ccsr->svr);
> +	svr = fsl_guts_read(&soc.dcfg_ccsr->svr);
>  	of_node_put(np);
>  
>  	/* Register soc device */

[Severity: High]
Does this code trigger a NULL pointer dereference on the error path if memory
allocation for soc_dev_attr fails?

In fsl_guts_init(), if the kzalloc_obj() allocation for soc_dev_attr fails
just below this block, the code jumps to err_nomem and falls through to the
err label:

fsl_guts_init() {
    ...
	soc_dev_attr = kzalloc_obj(*soc_dev_attr);
	if (!soc_dev_attr)
		goto err_nomem;
    ...
err_nomem:
	ret = -ENOMEM;
err:
	kfree(soc_dev_attr->family);
	kfree(soc_dev_attr->soc_id);
    ...
}

Since soc_dev_attr is NULL when taking this path, won't calling
kfree(soc_dev_attr->family) dereference the NULL pointer during device
initialization?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=3

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 4/8] soc: fsl: guts: make it easier to determine on which SoC we are running

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Dereferencing a NULL pointer on the `soc_dev_attr` allocation error path in `fsl_guts_init()`.
--

--- Patch [4]: [PATCH v3 phy-next 4/8] soc: fsl: guts: make it easier to determine on which SoC we are running ---
commit 27fff30cda7c763b92c9fd890450416b6a304140
Author: Ioana Ciornei <ioana.ciornei@nxp.com>

soc: fsl: guts: make it easier to determine on which SoC we are running

The guts driver will need to easily determine on which SoC it's running
when it will need to perform RCW override at runtime. The guts driver
knows this already because fsl_guts_init() reads the QorIQ/Layerscape
architectural System Version Register (SVR), but it doesn't save this
for later lookups.

> diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
> index e1b3a054e681b..922560d987824 100644
> --- a/drivers/soc/fsl/guts.c
> +++ b/drivers/soc/fsl/guts.c
[ ... ]
> @@ -239,12 +274,12 @@ static int __init fsl_guts_init(void)
>  	if (!soc_dev_attr->family)
>  		goto err_nomem;
>  
> -	soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "svr:0x%08x", svr);
> +	soc_dev_attr->soc_id = kasprintf(GFP_KERNEL, "svr:0x%08x", soc.svr);
>  	if (!soc_dev_attr->soc_id)
>  		goto err_nomem;

[Severity: High]
Does jumping to err_nomem when soc_dev_attr allocation fails cause a NULL
pointer dereference?

In an earlier commit in this series ("soc: fsl: guts: add a global structure
to hold state"), the error path in fsl_guts_init() was modified to unmap
soc.dcfg_ccsr, but it seems the cleanup assumes soc_dev_attr is valid:

fsl_guts_init()
	soc_dev_attr = kzalloc_obj(*soc_dev_attr);
	if (!soc_dev_attr)
		goto err_nomem;
    ...
err_nomem:
	ret = -ENOMEM;
err:
	kfree(soc_dev_attr->family);

If the dynamic allocation for soc_dev_attr fails, wouldn't we dereference
the NULL pointer when freeing its members in the err block?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=4

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 6/8] dt-bindings: fsl: layerscape-dcfg: define DCFG_DCSR region

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Dereferencing NULL pointer `soc_dev_attr` on the error path.
--

--- Patch [6]: [PATCH v3 phy-next 6/8] dt-bindings: fsl: layerscape-dcfg: define DCFG_DCSR region ---
commit 87aa71bb5ea85cc0388e2ad87f720397de246b0d
Author: Vladimir Oltean <vladimir.oltean@nxp.com>

dt-bindings: fsl: layerscape-dcfg: define DCFG_DCSR region

This commit defines a secondary "reg" array element in the DCFG binding
for the hidden DCSR register space. Exposing this area permits operating
systems to perform Reset Configuration Word (RCW) overrides, enabling
runtime reconfiguration of networking SerDes lanes.

> diff --git a/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml b/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml
> index 3fb0534ea597b..fc14fd0bf84b1 100644
> --- a/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml
> +++ b/Documentation/devicetree/bindings/soc/fsl/fsl,layerscape-dcfg.yaml
> @@ -36,7 +36,20 @@ properties:
>            - const: simple-mfd
>  
>    reg:
> -    maxItems: 1
> +    minItems: 1

[ ... ]

[Severity: High]
Does the error path in fsl_guts_init() cause a NULL pointer dereference?

drivers/soc/fsl/guts.c:fsl_guts_init() {
    ...
	soc.dcfg_ccsr = of_iomap(np, DCFG_CCSR);
	if (!soc.dcfg_ccsr) {
		of_node_put(np);
		goto err_nomem;
	}
    ...
err_nomem:
	ret = -ENOMEM;
err:
	kfree(soc_dev_attr->family);
	kfree(soc_dev_attr->soc_id);
    ...
}

If of_iomap() or the allocation for soc_dev_attr fails in fsl_guts_init(),
the error handling path jumps to err_nomem while soc_dev_attr is still NULL.

The cleanup code at the err label unconditionally evaluates
soc_dev_attr->family and soc_dev_attr->soc_id before passing them to
kfree(). Reading these fields dereferences the NULL soc_dev_attr pointer,
which can cause an immediate page fault during initialization.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=6

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 5/8] soc: fsl: guts: make fsl_soc_data available after fsl_guts_init()

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] NULL pointer dereference in the error handling path of `fsl_guts_init()`.
--

--- Patch [5]: [PATCH v3 phy-next 5/8] soc: fsl: guts: make fsl_soc_data available after fsl_guts_init() ---
commit d316b69e1c587b56ad6f914ddbd2e412490c8594
Author: Vladimir Oltean <vladimir.oltean@nxp.com>
soc: fsl: guts: make fsl_soc_data available after fsl_guts_init()

This commit preserves the soc_data determined during fsl_guts_init() so it
can be referenced later for RCW override methods. It updates the error
handling paths to properly clean up the saved state if initialization fails.

> diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
> --- a/drivers/soc/fsl/guts.c
> +++ b/drivers/soc/fsl/guts.c
> @@ -231,10 +232,9 @@ static int __init fsl_guts_init(void)
>  {
> -	struct soc_device_attribute *soc_dev_attr;
> +	struct soc_device_attribute *soc_dev_attr = NULL;
>  	static struct soc_device *soc_dev;

[ ... ]

> @@ -243,12 +243,12 @@ static int __init fsl_guts_init(void)
>  	np = of_find_matching_node_and_match(NULL, fsl_guts_of_match, &match);
>  	if (!np)
>  		return 0;
> -	soc_data = match->data;
> +	soc.data = match->data;
>  
>  	soc.dcfg_ccsr = of_iomap(np, DCFG_CCSR);
>  	if (!soc.dcfg_ccsr) {
>  		of_node_put(np);
> -		return -ENOMEM;
> +		goto err_nomem;
>  	}

[Severity: High]
If of_iomap() fails here, the code executes a goto to the error path while
soc_dev_attr is still explicitly NULL. 

Down in the error cleanup path, the code unconditionally dereferences 
soc_dev_attr to retrieve its members:

> @@ -311,8 +311,11 @@ static int __init fsl_guts_init(void)
>  	kfree(soc_dev_attr->revision);
>  	kfree(soc_dev_attr->serial_number);
>  	kfree(soc_dev_attr);

Will this trigger a NULL pointer dereference?

While kfree() itself handles NULL pointers gracefully, evaluating
soc_dev_attr->revision and soc_dev_attr->serial_number requires reading
memory from the NULL soc_dev_attr pointer before passing the argument to
kfree().

> -	iounmap(soc.dcfg_ccsr);
> -	soc.dcfg_ccsr = NULL;
> +	if (soc.dcfg_ccsr) {
> +		iounmap(soc.dcfg_ccsr);
> +		soc.dcfg_ccsr = NULL;
> +	}
> +	soc.data = NULL;
>  
>  	return ret;
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=5

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state

[Vladimir Oltean]

On Mon, Jul 20, 2026 at 01:50:49PM +0000, sashiko-bot@kernel.org wrote:
> [Severity: High]
> Will jumping to err_nomem cause a NULL pointer dereference?
> 
> If the allocation of soc_dev_attr fails, the code now jumps to the error
> handling path which dereferences soc_dev_attr to free its members:
> 
> > @@ -269,6 +271,8 @@ static int __init fsl_guts_init(void)
> >  	kfree(soc_dev_attr->revision);
> >  	kfree(soc_dev_attr->serial_number);
> 
> Since soc_dev_attr is NULL when this allocation fails, won't evaluating
> soc_dev_attr->revision and other members result in a kernel panic?

yeah.

I tried to be nice and play with the existing "err_nomem" and "err"
error handling labels, but the truth is that a single error cleanup path
is just an antipattern that doesn't scale as code becomes more complex.

Instead of having a single error handling procedure which needs to
carefully fend off from various invalid contexts from all the goto
sites, the standard convention is to have an incremental set of labels
which undo just the setup that was done up to the goto jump site.

I'll refactor fsl_guts_init() to use that convention for v4.

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 7/8] soc: fsl: guts: implement the RCW override procedure

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 6 potential issue(s) to consider:

New issues:
- [High] SerDes block index validation incorrectly uses `>=` instead of `>`, rejecting the highest valid SerDes index and breaking configuration.
- [High] SoC-specific GENMASK macros execute undefined behavior via negative shifts when `lane >= 4`.
- [High] Bulk RCW override initialization clobbers active configurations on uninitialized lanes.
- [Medium] Non-atomic read-modify-write on shared state causes lost lane initializations.
- [Medium] Extremely short 1 microsecond timeout for hardware register state reflection causes spurious failures.

Pre-existing issues:
- [High] Dereferencing `soc_dev_attr` on the error path causes a kernel panic.
--

--- Patch [7]: [PATCH v3 phy-next 7/8] soc: fsl: guts: implement the RCW override procedure ---
commit a1f1c62bf1e452bc97a7f348215d8401bd095833
Author: Ioana Ciornei <ioana.ciornei@nxp.com>
soc: fsl: guts: implement the RCW override procedure

This commit adds support for the RCW override procedure to enable runtime
reconfiguration of SerDes lane protocols through the DCFG DCSR space.
It exports functions to notify boot time modes, validate protocol changes,
and execute the RCW procedure on specific SoCs.

> diff --git a/drivers/soc/fsl/guts.c b/drivers/soc/fsl/guts.c
> index b9973ca3a440d..4d49d1b953007 100644
> --- a/drivers/soc/fsl/guts.c
> +++ b/drivers/soc/fsl/guts.c
[ ... ]
> @@ -210,6 +210,13 @@ static int fsl_guts_rcw_rmw(int offset, u32 val, u32 mask)
>  {
>  	u32 rcwcr, rcwsr = fsl_guts_read(&soc.dcfg_ccsr->rcwsr[offset]);
>  
>  	rcwcr = rcwsr & ~mask;
>  	rcwcr |= val;
>  	fsl_guts_write(&soc.dcfg_dcsr->rcwcr[offset], rcwcr);
>  
>  	/* Updates to RCWCR should be visible back in RCWSR */
>  	return read_poll_timeout_atomic(fsl_guts_read, rcwsr, rcwsr == rcwcr,
>  					0, RCW_TIMEOUT_US, false,
>  					&soc.dcfg_ccsr->rcwsr[offset]);
>  }

[Severity: Medium]
Is a 1 microsecond timeout sufficient for cross-domain hardware reflections?

The atomic polling loop strictly uses RCW_TIMEOUT_US set to 1, which might be
too brief for the hardware to propagate the state across clock domains. This
could lead to spurious -ETIMEDOUT failures when configuring SerDes lane modes.

[ ... ]
> +static int
> +fsl_guts_serdes_get_rcw_override(int serdes_idx, int lane,
> +				 enum lynx_lane_mode lane_mode,
> +				 struct fsl_soc_serdes_rcw_override *override)
> +{
> +	const struct fsl_soc_data *soc_data = soc.data;
> +
> +	if (!soc_data)
> +		return -ENODEV;
> +
> +	if (serdes_idx >= soc_data->num_serdes || serdes_idx <= 0)
> +		return -ERANGE;

[Severity: High]
Does this bounds check incorrectly reject the highest valid SerDes index?

If soc_data->num_serdes is 2, a client driver attempting to initialize
serdes_idx 2 will trigger this return path because 2 >= 2 evaluates to true.
Should this use > instead of >= to allow configuring the last SerDes block?

> +
> +	if (lane >= MAX_NUM_LANES || lane < 0)
> +		return -ERANGE;

[Severity: High]
Can this loose bounds check lead to undefined behavior in the SoC-specific
macros?

MAX_NUM_LANES is 8, but macros like LS1088A_RCWSR30_SRDS_PRTCL_S2_LNn
assume a maximum of 4 lanes:

GENMASK(3 + 4 * (3 - (lane)), 4 * (3 - (lane)))

If a PHY driver requests a lane index of 4 or higher on LS1088A, this macro
evaluates to GENMASK(-1, -4), which causes negative bitwise shifts.

[ ... ]
> +int fsl_guts_lane_init(int serdes_idx, int lane, enum lynx_lane_mode lane_mode)
> +{
> +	int err;
> +
> +	err = fsl_guts_lane_validate(serdes_idx, lane, lane_mode);
> +	if (err)
> +		return err;
> +
> +	soc.lane_mode[serdes_idx - 1][lane] = lane_mode;
> +	soc.lanes_initialized[serdes_idx - 1] |= BIT(lane);

[Severity: Medium]
Could concurrent PHY probes cause lost lane initialization status here?

This performs a non-atomic read-modify-write on the shared global state array
without holding a lock. If two PHY instances on the same SerDes controller
probe asynchronously, one update to soc.lanes_initialized might overwrite the
other.

[ ... ]
> +static int ls2088a_serdes_init_rcwcr(int serdes_idx)
> +{
> +	int i, err;
> +	u32 reg;
[ ... ]
> +	/* We need to configure the initial state of all lanes for
> +	 * the SerDes block #1
> +	 */
> +	for_each_set_bit(i, &soc.lanes_initialized[serdes_idx - 1], MAX_NUM_LANES)
> +		if (lynx_lane_mode_uses_xgmii_mac(soc.lane_mode[serdes_idx - 1][i]))
> +			reg |= LS2088A_RCWSR30_SRDS_CLK_SEL_XGMII_Ln_S1(i);
> +
> +	err = fsl_guts_rcw_rmw(29, reg,
> +			       LS2088A_RCWSR30_SRDS_CLK_EN_SEL_XGMII_S1 |
> +			       LS2088A_RCWSR30_SRDS_CLK_SEL_MSK);

[Severity: High]
Does this bulk RCW override initialization clobber active configurations on
uninitialized lanes?

If a client driver for lane A calls fsl_guts_lane_set_mode() before sibling
lane B has initialized, this read-modify-write uses a mask spanning all
lanes (LS2088A_RCWSR30_SRDS_CLK_SEL_MSK). This forcefully writes zeroes
to lane B's configuration because it relies purely on the software state
in soc.lanes_initialized, potentially discarding its bootloader-configured
active state.

[ ... ]
>  static int __init fsl_guts_init(void)
>  {
>  	struct soc_device_attribute *soc_dev_attr = NULL;
[ ... ]
>  err_nomem:
>  	ret = -ENOMEM;
>  err:
>  	kfree(soc_dev_attr->family);
>  	kfree(soc_dev_attr->soc_id);
>  	kfree(soc_dev_attr->revision);
>  	kfree(soc_dev_attr->serial_number);
>  	kfree(soc_dev_attr);

[Severity: High]
This is a pre-existing issue, but can dereferencing soc_dev_attr here cause
a kernel panic on the error path?

If earlier initialization steps fail, such as of_iomap() failing and jumping
to err_nomem before soc_dev_attr is allocated, soc_dev_attr will remain NULL.
The unconditional access to member variables like soc_dev_attr->family will
result in a NULL pointer dereference.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=7

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 8/8] phy: lynx-10g: use RCW override procedure for dynamic protocol change

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] An off-by-one error in the bounds check for `serdes_idx` within `fsl_guts_serdes_get_rcw_override` incorrectly rejects the final SerDes block on the SoC.
- [High] Uninitialized lanes on `ls2088a` SerDes1 have their XGMII clock selection bits blindly overwritten to GMII during RCW initialization.
--

--- Patch [8]: [PATCH v3 phy-next 8/8] phy: lynx-10g: use RCW override procedure for dynamic protocol change ---
commit 7e0651dd0ae8f0fcd028d4877a54ccdb1dcfef2d
Author: Vladimir Oltean <vladimir.oltean@nxp.com>

phy: lynx-10g: use RCW override procedure for dynamic protocol change

Up until this patch, the only protocol change supported was between
1000Base-X/SGMII and 2500Base-X. The others require an RCW override
procedure which was lacking.

Since now the guts driver provides the means of applying this procedure,
make use of it and remove any comment which mentioned the limitation.

> diff --git a/drivers/phy/freescale/phy-fsl-lynx-10g.c b/drivers/phy/freescale/phy-fsl-lynx-10g.c
> index 38def160ef1aa..5ece7889aed7b 100644
> --- a/drivers/phy/freescale/phy-fsl-lynx-10g.c
> +++ b/drivers/phy/freescale/phy-fsl-lynx-10g.c
[ ... ]
> @@ -1225,6 +1223,13 @@ static int lynx_10g_set_mode(struct phy *phy, enum phy_mode mode, int submode)
>  	if (powered_up)
>  		lynx_10g_lane_halt(phy);
>  
> +	if (lynx_10g_lane_mode_needs_rcw_override(lane, lane_mode)) {
> +		err = fsl_guts_lane_set_mode(priv->info->index, lane->id,
> +					     lane_mode);

[Severity: High]
This isn't a bug introduced by this patch, but does the underlying RCW override
procedure contain an off-by-one error that rejects the final SerDes block?

When this calls into fsl_guts_lane_set_mode(), it eventually hits a bounds
check in fsl_guts_serdes_get_rcw_override():

drivers/soc/fsl/guts.c:fsl_guts_serdes_get_rcw_override() {
    ...
    if (serdes_idx >= soc_data->num_serdes || serdes_idx <= 0)
        return -ERANGE;
    ...
}

Since serdes_idx is 1-based, and this uses >= instead of >, won't this
completely fail to apply the dynamic protocol change for any interface on
the last SerDes block (e.g. index 2 on SoCs with 2 SerDes blocks)?

[Severity: High]
This is a pre-existing issue, but does initializing the SerDes RCW
inadvertently corrupt unmanaged lanes on LS2088A?

When fsl_guts_lane_set_mode() is called for the first time on an LS2088A SoC,
it triggers ls2088a_serdes_init_rcwcr() to build a new state for the RCW
utilizing soc.lanes_initialized:

drivers/soc/fsl/guts.c:ls2088a_serdes_init_rcwcr() {
    ...
    for_each_set_bit(i, &soc.lanes_initialized[serdes_idx - 1], MAX_NUM_LANES)
        if (lynx_lane_mode_uses_xgmii_mac(soc.lane_mode[serdes_idx - 1][i]))
            reg |= LS2088A_RCWSR30_SRDS_CLK_SEL_XGMII_Ln_S1(i);

    err = fsl_guts_rcw_rmw(29, reg,
                   LS2088A_RCWSR30_SRDS_CLK_EN_SEL_XGMII_S1 |
                   LS2088A_RCWSR30_SRDS_CLK_SEL_MSK);
    ...
}

Because the blanket mask LS2088A_RCWSR30_SRDS_CLK_SEL_MSK clears the bits
for all 8 lanes but only restores them for explicitly managed lanes, won't
unmanaged lanes (such as those used by DPDK or alternative firmware) have
their hardware protocol selection forced to 0 (GMII), potentially breaking
network connectivity?

> +		if (err)
> +			goto out;
> +	}
> +
>  	err = lynx_10g_lane_disable_pcvt(lane, lane->mode);
>  	if (err)
>  		goto out;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260720133642.136324-1-vladimir.oltean@nxp.com?part=8

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state

[Vladimir Oltean]

Hi Michael,

On Mon, Jul 20, 2026 at 05:01:39PM +0300, Vladimir Oltean wrote:
> On Mon, Jul 20, 2026 at 01:50:49PM +0000, sashiko-bot@kernel.org wrote:
> > [Severity: High]
> > Will jumping to err_nomem cause a NULL pointer dereference?
> > 
> > If the allocation of soc_dev_attr fails, the code now jumps to the error
> > handling path which dereferences soc_dev_attr to free its members:
> > 
> > > @@ -269,6 +271,8 @@ static int __init fsl_guts_init(void)
> > >  	kfree(soc_dev_attr->revision);
> > >  	kfree(soc_dev_attr->serial_number);
> > 
> > Since soc_dev_attr is NULL when this allocation fails, won't evaluating
> > soc_dev_attr->revision and other members result in a kernel panic?
> 
> yeah.
> 
> I tried to be nice and play with the existing "err_nomem" and "err"
> error handling labels, but the truth is that a single error cleanup path
> is just an antipattern that doesn't scale as code becomes more complex.
> 
> Instead of having a single error handling procedure which needs to
> carefully fend off from various invalid contexts from all the goto
> sites, the standard convention is to have an incremental set of labels
> which undo just the setup that was done up to the goto jump site.
> 
> I'll refactor fsl_guts_init() to use that convention for v4.

While I was doing this refactoring, I noticed that this code path:

	if (soc_data)
		soc_uid = fsl_guts_get_soc_uid(soc_data->sfp_compat,
					       soc_data->uid_offset);
	if (soc_uid) {
		soc_dev_attr->serial_number = kasprintf(GFP_KERNEL, "%016llX",
							soc_uid);
		// no kasprintf() NULL return code check here, unlike
		// the rest of the code
	}

	soc_dev = soc_device_register(soc_dev_attr);
	if (IS_ERR(soc_dev)) {
		ret = PTR_ERR(soc_dev);
		goto err;
	}

proceeds with soc_device_register() even if fsl_guts_get_soc_uid() was
able to get a soc unique id from the security fuse processor, but we
fail to print that string to the soc_dev_attr->serial_number variable.

I don't see anywhere mentioned in commit 786dde1e59d7 ("soc: fsl: guts:
add serial_number support") that this would be intentional, so it can
just as well be an omission.

Is it OK if I replace the silent failure with a loud failure of the
entire fsl_guts_init()? The code executes only if soc_uid is non-zero
anyway.

kasprintf() fails only for memory related reasons.

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 8/8] phy: lynx-10g: use RCW override procedure for dynamic protocol change

[Vinod Koul]

On 20-07-26, 16:36, Vladimir Oltean wrote:
> Up until this patch, the only protocol change supported was between
> 1000Base-X/SGMII and 2500Base-X. The others require an RCW override
> procedure which was lacking.
> 
> Since now the guts driver provides the means of applying this procedure,
> make use of it and remove any comment which mentioned the limitation.

lgtm, is there any dependency. If not I can pick it

-- 
~Vinod

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 8/8] phy: lynx-10g: use RCW override procedure for dynamic protocol change

[Vladimir Oltean]

Hi Vinod,

On Mon, Jul 20, 2026 at 10:04:23PM +0530, Vinod Koul wrote:
> On 20-07-26, 16:36, Vladimir Oltean wrote:
> > Up until this patch, the only protocol change supported was between
> > 1000Base-X/SGMII and 2500Base-X. The others require an RCW override
> > procedure which was lacking.
> > 
> > Since now the guts driver provides the means of applying this procedure,
> > make use of it and remove any comment which mentioned the limitation.
> 
> lgtm, is there any dependency. If not I can pick it
> 
> -- 
> ~Vinod

There is no other direct dependency, but I need to send a v4 tomorrow
due to some valid Sashiko comments.

There is also this patch which seems to have been forgotten:
https://patchwork.kernel.org/project/linux-phy/patch/20260612125731.133330-1-vladimir.oltean@nxp.com/

Thanks,
Vladimir

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state

[Michael Walle]

[-- Attachment #1.1: Type: text/plain, Size: 2871 bytes --]

Hi Vladimir,

On Mon Jul 20, 2026 at 4:57 PM CEST, Vladimir Oltean wrote:
> On Mon, Jul 20, 2026 at 05:01:39PM +0300, Vladimir Oltean wrote:
>> On Mon, Jul 20, 2026 at 01:50:49PM +0000, sashiko-bot@kernel.org wrote:
>> > [Severity: High]
>> > Will jumping to err_nomem cause a NULL pointer dereference?
>> > 
>> > If the allocation of soc_dev_attr fails, the code now jumps to the error
>> > handling path which dereferences soc_dev_attr to free its members:
>> > 
>> > > @@ -269,6 +271,8 @@ static int __init fsl_guts_init(void)
>> > >  	kfree(soc_dev_attr->revision);
>> > >  	kfree(soc_dev_attr->serial_number);
>> > 
>> > Since soc_dev_attr is NULL when this allocation fails, won't evaluating
>> > soc_dev_attr->revision and other members result in a kernel panic?
>> 
>> yeah.
>> 
>> I tried to be nice and play with the existing "err_nomem" and "err"
>> error handling labels, but the truth is that a single error cleanup path
>> is just an antipattern that doesn't scale as code becomes more complex.
>> 
>> Instead of having a single error handling procedure which needs to
>> carefully fend off from various invalid contexts from all the goto
>> sites, the standard convention is to have an incremental set of labels
>> which undo just the setup that was done up to the goto jump site.
>> 
>> I'll refactor fsl_guts_init() to use that convention for v4.
>
> While I was doing this refactoring, I noticed that this code path:
>
> 	if (soc_data)
> 		soc_uid = fsl_guts_get_soc_uid(soc_data->sfp_compat,
> 					       soc_data->uid_offset);
> 	if (soc_uid) {
> 		soc_dev_attr->serial_number = kasprintf(GFP_KERNEL, "%016llX",
> 							soc_uid);
> 		// no kasprintf() NULL return code check here, unlike
> 		// the rest of the code

It's a long time ago, but it seems to be designed be optional. In
that case a NULL pointer check was omitted, because nothing bad can
happen anyway. Either you'll free the string if != NULL or you don't
free anything.

> 	}
>
> 	soc_dev = soc_device_register(soc_dev_attr);
> 	if (IS_ERR(soc_dev)) {
> 		ret = PTR_ERR(soc_dev);
> 		goto err;
> 	}
>
> proceeds with soc_device_register() even if fsl_guts_get_soc_uid() was
> able to get a soc unique id from the security fuse processor, but we
> fail to print that string to the soc_dev_attr->serial_number variable.
>
> I don't see anywhere mentioned in commit 786dde1e59d7 ("soc: fsl: guts:
> add serial_number support") that this would be intentional, so it can
> just as well be an omission.
>
> Is it OK if I replace the silent failure with a loud failure of the
> entire fsl_guts_init()? The code executes only if soc_uid is non-zero
> anyway.
>
> kasprintf() fails only for memory related reasons.

I don't have a strong opinion. If you ever get that error, something
is seriously bad anyway.

-michael

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 297 bytes --]

[-- Attachment #2: Type: text/plain, Size: 112 bytes --]

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy

Re: [PATCH v3 phy-next 2/8] soc: fsl: guts: add a global structure to hold state

[Vladimir Oltean]

Hi Michael,

On Tue, Jul 21, 2026 at 10:44:30AM +0200, Michael Walle wrote:
> Hi Vladimir,
> 
> On Mon Jul 20, 2026 at 4:57 PM CEST, Vladimir Oltean wrote:
> > On Mon, Jul 20, 2026 at 05:01:39PM +0300, Vladimir Oltean wrote:
> >> On Mon, Jul 20, 2026 at 01:50:49PM +0000, sashiko-bot@kernel.org wrote:
> >> > [Severity: High]
> >> > Will jumping to err_nomem cause a NULL pointer dereference?
> >> > 
> >> > If the allocation of soc_dev_attr fails, the code now jumps to the error
> >> > handling path which dereferences soc_dev_attr to free its members:
> >> > 
> >> > > @@ -269,6 +271,8 @@ static int __init fsl_guts_init(void)
> >> > >  	kfree(soc_dev_attr->revision);
> >> > >  	kfree(soc_dev_attr->serial_number);
> >> > 
> >> > Since soc_dev_attr is NULL when this allocation fails, won't evaluating
> >> > soc_dev_attr->revision and other members result in a kernel panic?
> >> 
> >> yeah.
> >> 
> >> I tried to be nice and play with the existing "err_nomem" and "err"
> >> error handling labels, but the truth is that a single error cleanup path
> >> is just an antipattern that doesn't scale as code becomes more complex.
> >> 
> >> Instead of having a single error handling procedure which needs to
> >> carefully fend off from various invalid contexts from all the goto
> >> sites, the standard convention is to have an incremental set of labels
> >> which undo just the setup that was done up to the goto jump site.
> >> 
> >> I'll refactor fsl_guts_init() to use that convention for v4.
> >
> > While I was doing this refactoring, I noticed that this code path:
> >
> > 	if (soc_data)
> > 		soc_uid = fsl_guts_get_soc_uid(soc_data->sfp_compat,
> > 					       soc_data->uid_offset);
> > 	if (soc_uid) {
> > 		soc_dev_attr->serial_number = kasprintf(GFP_KERNEL, "%016llX",
> > 							soc_uid);
> > 		// no kasprintf() NULL return code check here, unlike
> > 		// the rest of the code
> 
> It's a long time ago, but it seems to be designed be optional. In
> that case a NULL pointer check was omitted, because nothing bad can
> happen anyway. Either you'll free the string if != NULL or you don't
> free anything.
> 
> > 	}
> >
> > 	soc_dev = soc_device_register(soc_dev_attr);
> > 	if (IS_ERR(soc_dev)) {
> > 		ret = PTR_ERR(soc_dev);
> > 		goto err;
> > 	}
> >
> > proceeds with soc_device_register() even if fsl_guts_get_soc_uid() was
> > able to get a soc unique id from the security fuse processor, but we
> > fail to print that string to the soc_dev_attr->serial_number variable.
> >
> > I don't see anywhere mentioned in commit 786dde1e59d7 ("soc: fsl: guts:
> > add serial_number support") that this would be intentional, so it can
> > just as well be an omission.
> >
> > Is it OK if I replace the silent failure with a loud failure of the
> > entire fsl_guts_init()? The code executes only if soc_uid is non-zero
> > anyway.
> >
> > kasprintf() fails only for memory related reasons.
> 
> I don't have a strong opinion. If you ever get that error, something
> is seriously bad anyway.

Thanks for responding.

I think the optionality is already taken care of by the "if (soc_uid)" check.
My intuition is that if the guts driver decides to try and populate the
soc_dev_attr->serial_number attribute, it should try and run this to
completion.

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy