[PATCH v2 net-next 0/2] net: enetc: improve the interface for obtaining phc_index

[PATCH v2 net-next 0/2] net: enetc: improve the interface for obtaining phc_index

[Wei Fang]

The first patch is to fix the issue that a sleeping function is called
in the context of rcu_read_lock(). The second patch is to use the generic
API instead of the custom API to get phc_index. In addition, the second
patch depends on the first patch to work.

Note, the first patch is a fix, but the issue does not exist in the net
tree, so the target tree is net-next.

---
v2 changes:
1. Add patch 1 to fix the issue that sleeping function called from
   invalid context
2. Fix the build warning of patch 2.
v1 link: https://lore.kernel.org/netdev/20250918074454.1742328-1-wei.fang@nxp.com/
---

Wei Fang (2):
  net: enetc: fix sleeping function called from rcu_read_lock() context
  net: enetc: use generic interfaces to get phc_index for ENETC v1

 drivers/net/ethernet/freescale/enetc/enetc.h  |  3 --
 .../ethernet/freescale/enetc/enetc_ethtool.c  | 29 ++++++++-----------
 .../net/ethernet/freescale/enetc/enetc_ptp.c  |  5 ----
 3 files changed, 12 insertions(+), 25 deletions(-)

-- 
2.34.1

[PATCH v2 net-next 1/2] net: enetc: fix sleeping function called from rcu_read_lock() context

[Wei Fang]

The rcu_read_lock() has been introduced in __ethtool_get_ts_info() since
the commit 4c61d809cf60 ("net: ethtool: Fix suspicious rcu_dereference
usage"). Therefore, the device drivers cannot use any sleeping functions
when implementing the callback of ethtool_ops::get_ts_info(). Currently,
pci_get_slot() is used in enetc_get_ts_info(), but it calls down_read()
which might sleep, so this is a potential issue. Therefore, to fix this
issue, pci_get_domain_bus_and_slot() is used to replace pci_get_slot()
in enetc_get_ts_info().

Reported-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Closes: https://lore.kernel.org/netdev/20250918124823.t3xlzn7w2glzkhnx@skbuf/
Fixes: f5b9a1cde0a2 ("net: enetc: add PTP synchronization support for ENETC v4")
Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc_ethtool.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c b/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
index 6215e9c68fc5..445bfd032e0f 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
@@ -895,7 +895,8 @@ static int enetc4_get_phc_index_by_pdev(struct enetc_si *si)
 		return -1;
 	}
 
-	timer_pdev = pci_get_slot(bus, devfn);
+	timer_pdev = pci_get_domain_bus_and_slot(pci_domain_nr(bus),
+						 bus->number, devfn);
 	if (!timer_pdev)
 		return -1;
 
-- 
2.34.1

[PATCH v2 net-next 2/2] net: enetc: use generic interfaces to get phc_index for ENETC v1

[Wei Fang]

The commit 61f132ca8c46 ("ptp: add helpers to get the phc_index by
of_node or dev") has added two generic interfaces to get the phc_index
of the PTP clock. This eliminates the need for PTP device drivers to
provide custom APIs for consumers to retrieve the phc_index. This has
already been implemented for ENETC v4 and is also applicable to ENETC
v1. Therefore, the global variable enetc_phc_index is removed from the
driver. ENETC v1 now uses the same interface as v4 to get phc_index.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc.h  |  3 ---
 .../ethernet/freescale/enetc/enetc_ethtool.c  | 26 +++++++------------
 .../net/ethernet/freescale/enetc/enetc_ptp.c  |  5 ----
 3 files changed, 10 insertions(+), 24 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.h b/drivers/net/ethernet/freescale/enetc/enetc.h
index 815afdc2ec23..0ec010a7d640 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc.h
@@ -493,9 +493,6 @@ struct enetc_msg_cmd_set_primary_mac {
 
 #define ENETC_CBDR_TIMEOUT	1000 /* usecs */
 
-/* PTP driver exports */
-extern int enetc_phc_index;
-
 /* SI common */
 u32 enetc_port_mac_rd(struct enetc_si *si, u32 reg);
 void enetc_port_mac_wr(struct enetc_si *si, u32 reg, u32 val);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c b/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
index 445bfd032e0f..71d052de669a 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
@@ -880,7 +880,7 @@ static int enetc_set_coalesce(struct net_device *ndev,
 	return 0;
 }
 
-static int enetc4_get_phc_index_by_pdev(struct enetc_si *si)
+static int enetc_get_phc_index_by_pdev(struct enetc_si *si)
 {
 	struct pci_bus *bus = si->pdev->bus;
 	struct pci_dev *timer_pdev;
@@ -888,6 +888,9 @@ static int enetc4_get_phc_index_by_pdev(struct enetc_si *si)
 	int phc_index;
 
 	switch (si->revision) {
+	case ENETC_REV_1_0:
+		devfn = PCI_DEVFN(0, 4);
+		break;
 	case ENETC_REV_4_1:
 		devfn = PCI_DEVFN(24, 0);
 		break;
@@ -906,18 +909,18 @@ static int enetc4_get_phc_index_by_pdev(struct enetc_si *si)
 	return phc_index;
 }
 
-static int enetc4_get_phc_index(struct enetc_si *si)
+static int enetc_get_phc_index(struct enetc_si *si)
 {
 	struct device_node *np = si->pdev->dev.of_node;
 	struct device_node *timer_np;
 	int phc_index;
 
 	if (!np)
-		return enetc4_get_phc_index_by_pdev(si);
+		return enetc_get_phc_index_by_pdev(si);
 
 	timer_np = of_parse_phandle(np, "ptp-timer", 0);
 	if (!timer_np)
-		return enetc4_get_phc_index_by_pdev(si);
+		return enetc_get_phc_index_by_pdev(si);
 
 	phc_index = ptp_clock_index_by_of_node(timer_np);
 	of_node_put(timer_np);
@@ -950,22 +953,13 @@ static int enetc_get_ts_info(struct net_device *ndev,
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
 	struct enetc_si *si = priv->si;
-	int *phc_idx;
 
 	if (!enetc_ptp_clock_is_enabled(si))
 		goto timestamp_tx_sw;
 
-	if (is_enetc_rev1(si)) {
-		phc_idx = symbol_get(enetc_phc_index);
-		if (phc_idx) {
-			info->phc_index = *phc_idx;
-			symbol_put(enetc_phc_index);
-		}
-	} else {
-		info->phc_index = enetc4_get_phc_index(si);
-		if (info->phc_index < 0)
-			goto timestamp_tx_sw;
-	}
+	info->phc_index = enetc_get_phc_index(si);
+	if (info->phc_index < 0)
+		goto timestamp_tx_sw;
 
 	enetc_get_ts_generic_info(ndev, info);
 
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_ptp.c b/drivers/net/ethernet/freescale/enetc/enetc_ptp.c
index 5243fc031058..b8413d3b4f16 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_ptp.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_ptp.c
@@ -7,9 +7,6 @@
 
 #include "enetc.h"
 
-int enetc_phc_index = -1;
-EXPORT_SYMBOL_GPL(enetc_phc_index);
-
 static struct ptp_clock_info enetc_ptp_caps = {
 	.owner		= THIS_MODULE,
 	.name		= "ENETC PTP clock",
@@ -92,7 +89,6 @@ static int enetc_ptp_probe(struct pci_dev *pdev,
 	if (err)
 		goto err_no_clock;
 
-	enetc_phc_index = ptp_qoriq->phc_index;
 	pci_set_drvdata(pdev, ptp_qoriq);
 
 	return 0;
@@ -118,7 +114,6 @@ static void enetc_ptp_remove(struct pci_dev *pdev)
 {
 	struct ptp_qoriq *ptp_qoriq = pci_get_drvdata(pdev);
 
-	enetc_phc_index = -1;
 	ptp_qoriq_free(ptp_qoriq);
 	pci_free_irq_vectors(pdev);
 	kfree(ptp_qoriq);
-- 
2.34.1

Re: [PATCH v2 net-next 0/2] net: enetc: improve the interface for obtaining phc_index

[Vladimir Oltean]

On Fri, Sep 19, 2025 at 04:45:07PM +0800, Wei Fang wrote:
> The first patch is to fix the issue that a sleeping function is called
> in the context of rcu_read_lock(). The second patch is to use the generic
> API instead of the custom API to get phc_index. In addition, the second
> patch depends on the first patch to work.
> 
> Note, the first patch is a fix, but the issue does not exist in the net
> tree, so the target tree is net-next.
> 
> ---
> v2 changes:
> 1. Add patch 1 to fix the issue that sleeping function called from
>    invalid context
> 2. Fix the build warning of patch 2.
> v1 link: https://lore.kernel.org/netdev/20250918074454.1742328-1-wei.fang@nxp.com/
> ---
> 
> Wei Fang (2):
>   net: enetc: fix sleeping function called from rcu_read_lock() context
>   net: enetc: use generic interfaces to get phc_index for ENETC v1
> 
>  drivers/net/ethernet/freescale/enetc/enetc.h  |  3 --
>  .../ethernet/freescale/enetc/enetc_ethtool.c  | 29 ++++++++-----------
>  .../net/ethernet/freescale/enetc/enetc_ptp.c  |  5 ----
>  3 files changed, 12 insertions(+), 25 deletions(-)
> 
> -- 
> 2.34.1
>

For the set:

Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>

Re: [PATCH v2 net-next 0/2] net: enetc: improve the interface for obtaining phc_index

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Fri, 19 Sep 2025 16:45:07 +0800 you wrote:
> The first patch is to fix the issue that a sleeping function is called
> in the context of rcu_read_lock(). The second patch is to use the generic
> API instead of the custom API to get phc_index. In addition, the second
> patch depends on the first patch to work.
> 
> Note, the first patch is a fix, but the issue does not exist in the net
> tree, so the target tree is net-next.
> 
> [...]

Here is the summary with links:
  - [v2,net-next,1/2] net: enetc: fix sleeping function called from rcu_read_lock() context
    https://git.kernel.org/netdev/net-next/c/26644c90e8fb
  - [v2,net-next,2/2] net: enetc: use generic interfaces to get phc_index for ENETC v1
    https://git.kernel.org/netdev/net-next/c/ac0e650fde45

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