diff mbox series

[net-next,v5,01/13] net: phy: Introduce ethernet link topology representation

Message ID 20231221180047.1924733-2-maxime.chevallier@bootlin.com (mailing list archive)
State Accepted
Commit 02018c544ef113e980a2349eba89003d6f399d22
Delegated to: Netdev Maintainers
Headers show
Series Introduce PHY listing and link_topology tracking | expand

Checks

Context Check Description
netdev/series_format success Posting correctly formatted
netdev/tree_selection success Clearly marked for net-next, async
netdev/ynl success Generated files up to date; no warnings/errors; GEN HAS DIFF 2 files changed, 954 insertions(+);
netdev/fixes_present success Fixes tag not required for -next series
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 5206 this patch: 5206
netdev/cc_maintainers warning 3 maintainers not CCed: keescook@chromium.org ahmed.zaki@intel.com sd@queasysnail.net
netdev/build_clang fail Errors and warnings before: 229 this patch: 229
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/deprecated_api success None detected
netdev/check_selftest success No net selftest shell script
netdev/verify_fixes success No Fixes tag
netdev/build_allmodconfig_warn success Errors and warnings before: 5533 this patch: 5533
netdev/checkpatch warning WARNING: line length of 81 exceeds 80 columns
netdev/build_clang_rust success No Rust files in patch. Skipping build
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0

Commit Message

Maxime Chevallier Dec. 21, 2023, 6 p.m. UTC
Link topologies containing multiple network PHYs attached to the same
net_device can be found when using a PHY as a media converter for use
with an SFP connector, on which an SFP transceiver containing a PHY can
be used.

With the current model, the transceiver's PHY can't be used for
operations such as cable testing, timestamping, macsec offload, etc.

The reason being that most of the logic for these configuration, coming
from either ethtool netlink or ioctls tend to use netdev->phydev, which
in multi-phy systems will reference the PHY closest to the MAC.

Introduce a numbering scheme allowing to enumerate PHY devices that
belong to any netdev, which can in turn allow userspace to take more
precise decisions with regard to each PHY's configuration.

The numbering is maintained per-netdev, in a phy_device_list.
The numbering works similarly to a netdevice's ifindex, with
identifiers that are only recycled once INT_MAX has been reached.

This prevents races that could occur between PHY listing and SFP
transceiver removal/insertion.

The identifiers are assigned at phy_attach time, as the numbering
depends on the netdevice the phy is attached to.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
V5: - Dropped the ASSERT_RTNL()
    - Made the phy_link_topo_get_phy inline
V4: - Moved the phy_link_topo_init() code to an inline header function
    - Made the code build without phylib

V3: - Renamed to phy_link_topology
    - Added assertions for RTNL
    - Various cleanups of leftover, unused test code
    - Made the PHY index u32

 MAINTAINERS                            |  2 +
 drivers/net/phy/Makefile               |  2 +-
 drivers/net/phy/phy_device.c           |  7 +++
 drivers/net/phy/phy_link_topology.c    | 66 +++++++++++++++++++++++++
 include/linux/netdevice.h              |  4 +-
 include/linux/phy.h                    |  4 ++
 include/linux/phy_link_topology.h      | 67 ++++++++++++++++++++++++++
 include/linux/phy_link_topology_core.h | 19 ++++++++
 include/uapi/linux/ethtool.h           | 16 ++++++
 net/core/dev.c                         |  3 ++
 10 files changed, 188 insertions(+), 2 deletions(-)
 create mode 100644 drivers/net/phy/phy_link_topology.c
 create mode 100644 include/linux/phy_link_topology.h
 create mode 100644 include/linux/phy_link_topology_core.h

Comments

Jakub Kicinski Jan. 4, 2024, 11:12 p.m. UTC | #1
On Thu, 21 Dec 2023 19:00:34 +0100 Maxime Chevallier wrote:
> @@ -2441,6 +2442,7 @@ struct net_device {
>  #if IS_ENABLED(CONFIG_CGROUP_NET_PRIO)
>  	struct netprio_map __rcu *priomap;
>  #endif
> +	struct phy_link_topology	link_topo;

Perhaps others would disagree but can we make this a pointer instead?
Only allocate it on demand, when first PHY gets attached?
Both saves space and netdevice.h will no longer need to know the
definition of the struct.

Complete noob question but I thought PHYs get attached at ndo_open
time for drivers, don't they? We shouldn't want to re-ID in that case.

>  	struct phy_device	*phydev;
>  	struct sfp_bus		*sfp_bus;
>  	struct lock_class_key	*qdisc_tx_busylock;

> @@ -10872,6 +10873,8 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name,
>  #ifdef CONFIG_NET_SCHED
>  	hash_init(dev->qdisc_hash);
>  #endif
> +	phy_link_topo_init(&dev->link_topo);
> +
>  	dev->priv_flags = IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM;
>  	setup(dev);
>  

I think you're missing a call to xa_destroy() somewhere, no?
Andrew Lunn Jan. 5, 2024, 2:21 a.m. UTC | #2
> Complete noob question but I thought PHYs get attached at ndo_open
> time for drivers, don't they? 

Some do. Others get attached during probe. And Copper PHYs inside SFPs
get attached when hot-plugged.

     Andrew
Maxime Chevallier Jan. 5, 2024, 9:29 a.m. UTC | #3
Hi Jakub,

On Thu, 4 Jan 2024 15:12:42 -0800
Jakub Kicinski <kuba@kernel.org> wrote:

> On Thu, 21 Dec 2023 19:00:34 +0100 Maxime Chevallier wrote:
> > @@ -2441,6 +2442,7 @@ struct net_device {
> >  #if IS_ENABLED(CONFIG_CGROUP_NET_PRIO)
> >  	struct netprio_map __rcu *priomap;
> >  #endif
> > +	struct phy_link_topology	link_topo;  
> 
> Perhaps others would disagree but can we make this a pointer instead?
> Only allocate it on demand, when first PHY gets attached?
> Both saves space and netdevice.h will no longer need to know the
> definition of the struct.
>
> Complete noob question but I thought PHYs get attached at ndo_open
> time for drivers, don't they? We shouldn't want to re-ID in that case.

I'll give it a try, it could be doable to try to keep the index for
_some_ PHYs. I had a first try at this in the very first iteration, but
that was lost when converting to xarray for the index management. Since
the phy_device keeps track of its own index, we can try to re-use it,
but I can see it becoming difficult if we destroy the topology when all
PHYs are detached (after ndo_close).

When re-creating the xarray at the next ndo_open, we would need to know
what was the previous next_index, and we can't know that until PHYs are
all re-attached.

For now it would work because in most cases, we only have one PHY
behind the MAC we can guarantee won't go away, but there are devices
out-there with 2 PHYs connected to the same MAC with a MUX in the
middle (that's one thing this series prepares for), and here it would
be more complex to re-allocate the topology while knowing what's the
next unused index. Hope that makes sense :)

To summarize, I don't think we can easily both dynamically
allocate/destroy the topology based only on the presence of at least
one PHY, and try to keep the PHY index the same for non-hot-pluggable
PHYs.

> 
> >  	struct phy_device	*phydev;
> >  	struct sfp_bus		*sfp_bus;
> >  	struct lock_class_key	*qdisc_tx_busylock;  
> 
> > @@ -10872,6 +10873,8 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name,
> >  #ifdef CONFIG_NET_SCHED
> >  	hash_init(dev->qdisc_hash);
> >  #endif
> > +	phy_link_topo_init(&dev->link_topo);
> > +
> >  	dev->priv_flags = IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM;
> >  	setup(dev);
> >    
> 
> I think you're missing a call to xa_destroy() somewhere, no?

Arg you're right... I was under the false assumption that because the
xarray wasn't dynamically allocated, we didn't need to perform any
cleanup, thanks for spotting this.

Maxime
Jakub Kicinski Jan. 5, 2024, 3:34 p.m. UTC | #4
On Fri, 5 Jan 2024 10:29:18 +0100 Maxime Chevallier wrote:
> I'll give it a try, it could be doable to try to keep the index for
> _some_ PHYs. I had a first try at this in the very first iteration, but
> that was lost when converting to xarray for the index management. Since
> the phy_device keeps track of its own index, we can try to re-use it,
> but I can see it becoming difficult if we destroy the topology when all
> PHYs are detached (after ndo_close).

What I was thinking is if phy already has phyindex, and that index is
unused in the xarray of netdev in question - try to use that phyindex
instead of allocating a new one?
IDK if phys can move between netdevs, or get freed/recreated
automatically - otherwise it should work? (If user unbinds and rebinds
the PHY driver for instance that's their fault, we can renumber).
diff mbox series

Patch

diff --git a/MAINTAINERS b/MAINTAINERS
index dda78b4ce707..f09b1d4e5487 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -7873,6 +7873,8 @@  F:	include/linux/mii.h
 F:	include/linux/of_net.h
 F:	include/linux/phy.h
 F:	include/linux/phy_fixed.h
+F:	include/linux/phy_link_topology.h
+F:	include/linux/phy_link_topology_core.h
 F:	include/linux/phylib_stubs.h
 F:	include/linux/platform_data/mdio-bcm-unimac.h
 F:	include/linux/platform_data/mdio-gpio.h
diff --git a/drivers/net/phy/Makefile b/drivers/net/phy/Makefile
index e35ea69d9cb4..a7a9640bfa3a 100644
--- a/drivers/net/phy/Makefile
+++ b/drivers/net/phy/Makefile
@@ -2,7 +2,7 @@ 
 # Makefile for Linux PHY drivers
 
 libphy-y			:= phy.o phy-c45.o phy-core.o phy_device.o \
-				   linkmode.o
+				   linkmode.o phy_link_topology.o
 mdio-bus-y			+= mdio_bus.o mdio_device.o
 
 ifdef CONFIG_MDIO_DEVICE
diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index 0c52a9eff188..d7c0812bd107 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c
@@ -29,6 +29,7 @@ 
 #include <linux/phy.h>
 #include <linux/phylib_stubs.h>
 #include <linux/phy_led_triggers.h>
+#include <linux/phy_link_topology.h>
 #include <linux/pse-pd/pse.h>
 #include <linux/property.h>
 #include <linux/rtnetlink.h>
@@ -1491,6 +1492,11 @@  int phy_attach_direct(struct net_device *dev, struct phy_device *phydev,
 
 		if (phydev->sfp_bus_attached)
 			dev->sfp_bus = phydev->sfp_bus;
+
+		err = phy_link_topo_add_phy(&dev->link_topo, phydev,
+					    PHY_UPSTREAM_MAC, dev);
+		if (err)
+			goto error;
 	}
 
 	/* Some Ethernet drivers try to connect to a PHY device before
@@ -1819,6 +1825,7 @@  void phy_detach(struct phy_device *phydev)
 	if (dev) {
 		phydev->attached_dev->phydev = NULL;
 		phydev->attached_dev = NULL;
+		phy_link_topo_del_phy(&dev->link_topo, phydev);
 	}
 	phydev->phylink = NULL;
 
diff --git a/drivers/net/phy/phy_link_topology.c b/drivers/net/phy/phy_link_topology.c
new file mode 100644
index 000000000000..34e7e08fbfc3
--- /dev/null
+++ b/drivers/net/phy/phy_link_topology.c
@@ -0,0 +1,66 @@ 
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Infrastructure to handle all PHY devices connected to a given netdev,
+ * either directly or indirectly attached.
+ *
+ * Copyright (c) 2023 Maxime Chevallier<maxime.chevallier@bootlin.com>
+ */
+
+#include <linux/phy_link_topology.h>
+#include <linux/netdevice.h>
+#include <linux/phy.h>
+#include <linux/rtnetlink.h>
+#include <linux/xarray.h>
+
+int phy_link_topo_add_phy(struct phy_link_topology *topo,
+			  struct phy_device *phy,
+			  enum phy_upstream upt, void *upstream)
+{
+	struct phy_device_node *pdn;
+	int ret;
+
+	pdn = kzalloc(sizeof(*pdn), GFP_KERNEL);
+	if (!pdn)
+		return -ENOMEM;
+
+	pdn->phy = phy;
+	switch (upt) {
+	case PHY_UPSTREAM_MAC:
+		pdn->upstream.netdev = (struct net_device *)upstream;
+		if (phy_on_sfp(phy))
+			pdn->parent_sfp_bus = pdn->upstream.netdev->sfp_bus;
+		break;
+	case PHY_UPSTREAM_PHY:
+		pdn->upstream.phydev = (struct phy_device *)upstream;
+		if (phy_on_sfp(phy))
+			pdn->parent_sfp_bus = pdn->upstream.phydev->sfp_bus;
+		break;
+	default:
+		ret = -EINVAL;
+		goto err;
+	}
+	pdn->upstream_type = upt;
+
+	ret = xa_alloc_cyclic(&topo->phys, &phy->phyindex, pdn, xa_limit_32b,
+			      &topo->next_phy_index, GFP_KERNEL);
+	if (ret)
+		goto err;
+
+	return 0;
+
+err:
+	kfree(pdn);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(phy_link_topo_add_phy);
+
+void phy_link_topo_del_phy(struct phy_link_topology *topo,
+			   struct phy_device *phy)
+{
+	struct phy_device_node *pdn = xa_erase(&topo->phys, phy->phyindex);
+
+	phy->phyindex = 0;
+
+	kfree(pdn);
+}
+EXPORT_SYMBOL_GPL(phy_link_topo_del_phy);
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 75c7725e5e4f..5baa5517f533 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -40,7 +40,6 @@ 
 #include <net/dcbnl.h>
 #endif
 #include <net/netprio_cgroup.h>
-
 #include <linux/netdev_features.h>
 #include <linux/neighbour.h>
 #include <uapi/linux/netdevice.h>
@@ -52,6 +51,7 @@ 
 #include <net/net_trackers.h>
 #include <net/net_debug.h>
 #include <net/dropreason-core.h>
+#include <linux/phy_link_topology_core.h>
 
 struct netpoll_info;
 struct device;
@@ -2047,6 +2047,7 @@  enum netdev_stat_type {
  *	@fcoe_ddp_xid:	Max exchange id for FCoE LRO by ddp
  *
  *	@priomap:	XXX: need comments on this one
+ *	@link_topo:	Physical link topology tracking attached PHYs
  *	@phydev:	Physical device may attach itself
  *			for hardware timestamping
  *	@sfp_bus:	attached &struct sfp_bus structure.
@@ -2441,6 +2442,7 @@  struct net_device {
 #if IS_ENABLED(CONFIG_CGROUP_NET_PRIO)
 	struct netprio_map __rcu *priomap;
 #endif
+	struct phy_link_topology	link_topo;
 	struct phy_device	*phydev;
 	struct sfp_bus		*sfp_bus;
 	struct lock_class_key	*qdisc_tx_busylock;
diff --git a/include/linux/phy.h b/include/linux/phy.h
index e9e85d347587..9f21eb380475 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -544,6 +544,9 @@  struct macsec_ops;
  * @drv: Pointer to the driver for this PHY instance
  * @devlink: Create a link between phy dev and mac dev, if the external phy
  *           used by current mac interface is managed by another mac interface.
+ * @phyindex: Unique id across the phy's parent tree of phys to address the PHY
+ *	      from userspace, similar to ifindex. A zero index means the PHY
+ *	      wasn't assigned an id yet.
  * @phy_id: UID for this device found during discovery
  * @c45_ids: 802.3-c45 Device Identifiers if is_c45.
  * @is_c45:  Set to true if this PHY uses clause 45 addressing.
@@ -643,6 +646,7 @@  struct phy_device {
 
 	struct device_link *devlink;
 
+	u32 phyindex;
 	u32 phy_id;
 
 	struct phy_c45_device_ids c45_ids;
diff --git a/include/linux/phy_link_topology.h b/include/linux/phy_link_topology.h
new file mode 100644
index 000000000000..91902263ec0e
--- /dev/null
+++ b/include/linux/phy_link_topology.h
@@ -0,0 +1,67 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * PHY device list allow maintaining a list of PHY devices that are
+ * part of a netdevice's link topology. PHYs can for example be chained,
+ * as is the case when using a PHY that exposes an SFP module, on which an
+ * SFP transceiver that embeds a PHY is connected.
+ *
+ * This list can then be used by userspace to leverage individual PHY
+ * capabilities.
+ */
+#ifndef __PHY_LINK_TOPOLOGY_H
+#define __PHY_LINK_TOPOLOGY_H
+
+#include <linux/ethtool.h>
+#include <linux/phy_link_topology_core.h>
+
+struct xarray;
+struct phy_device;
+struct net_device;
+struct sfp_bus;
+
+struct phy_device_node {
+	enum phy_upstream upstream_type;
+
+	union {
+		struct net_device	*netdev;
+		struct phy_device	*phydev;
+	} upstream;
+
+	struct sfp_bus *parent_sfp_bus;
+
+	struct phy_device *phy;
+};
+
+static inline struct phy_device *
+phy_link_topo_get_phy(struct phy_link_topology *topo, u32 phyindex)
+{
+	struct phy_device_node *pdn = xa_load(&topo->phys, phyindex);
+
+	if (pdn)
+		return pdn->phy;
+
+	return NULL;
+}
+
+#if IS_ENABLED(CONFIG_PHYLIB)
+int phy_link_topo_add_phy(struct phy_link_topology *topo,
+			  struct phy_device *phy,
+			  enum phy_upstream upt, void *upstream);
+
+void phy_link_topo_del_phy(struct phy_link_topology *lt, struct phy_device *phy);
+
+#else
+static inline int phy_link_topo_add_phy(struct phy_link_topology *topo,
+					struct phy_device *phy,
+					enum phy_upstream upt, void *upstream)
+{
+	return 0;
+}
+
+static inline void phy_link_topo_del_phy(struct phy_link_topology *topo,
+					 struct phy_device *phy)
+{
+}
+#endif
+
+#endif /* __PHY_LINK_TOPOLOGY_H */
diff --git a/include/linux/phy_link_topology_core.h b/include/linux/phy_link_topology_core.h
new file mode 100644
index 000000000000..78c75f909489
--- /dev/null
+++ b/include/linux/phy_link_topology_core.h
@@ -0,0 +1,19 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __PHY_LINK_TOPOLOGY_CORE_H
+#define __PHY_LINK_TOPOLOGY_CORE_H
+
+struct xarray;
+
+struct phy_link_topology {
+	struct xarray phys;
+
+	u32 next_phy_index;
+};
+
+static inline void phy_link_topo_init(struct phy_link_topology *topo)
+{
+	xa_init_flags(&topo->phys, XA_FLAGS_ALLOC1);
+	topo->next_phy_index = 1;
+}
+
+#endif /* __PHY_LINK_TOPOLOGY_CORE_H */
diff --git a/include/uapi/linux/ethtool.h b/include/uapi/linux/ethtool.h
index 0787d561ace0..9cff798c6df9 100644
--- a/include/uapi/linux/ethtool.h
+++ b/include/uapi/linux/ethtool.h
@@ -2216,4 +2216,20 @@  struct ethtool_link_settings {
 	 * __u32 map_lp_advertising[link_mode_masks_nwords];
 	 */
 };
+
+/**
+ * enum phy_upstream - Represents the upstream component a given PHY device
+ * is connected to, as in what is on the other end of the MII bus. Most PHYs
+ * will be attached to an Ethernet MAC controller, but in some cases, there's
+ * an intermediate PHY used as a media-converter, which will driver another
+ * MII interface as its output.
+ * @PHY_UPSTREAM_MAC: Upstream component is a MAC (a switch port,
+ *		      or ethernet controller)
+ * @PHY_UPSTREAM_PHY: Upstream component is a PHY (likely a media converter)
+ */
+enum phy_upstream {
+	PHY_UPSTREAM_MAC,
+	PHY_UPSTREAM_PHY,
+};
+
 #endif /* _UAPI_LINUX_ETHTOOL_H */
diff --git a/net/core/dev.c b/net/core/dev.c
index b87504078320..476acd22956c 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -153,6 +153,7 @@ 
 #include <linux/prandom.h>
 #include <linux/once_lite.h>
 #include <net/netdev_rx_queue.h>
+#include <linux/phy_link_topology_core.h>
 
 #include "dev.h"
 #include "net-sysfs.h"
@@ -10872,6 +10873,8 @@  struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name,
 #ifdef CONFIG_NET_SCHED
 	hash_init(dev->qdisc_hash);
 #endif
+	phy_link_topo_init(&dev->link_topo);
+
 	dev->priv_flags = IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM;
 	setup(dev);