kernel/git/next/linux-next.git/drivers/net/dsa, branch master

net: dsa: motorcomm: Add LED support

2026-09-03T02:58:54+00:00

Each port has at most 3 LEDs connected and can be declared using the standard LEDs structure. Currently, only parallel mode and strict 1:1 mapping (i.e. the HW default) are supported. Signed-off-by: David Yang Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20260831044818.4103477-5-mmyangfl@gmail.com Signed-off-by: Jakub Kicinski

net: dsa: motorcomm: Split SMI module

2026-09-03T02:58:54+00:00

SMI operations are going to be used across different modules. Minor changes are applied to fix build errors. Signed-off-by: David Yang Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20260831044818.4103477-4-mmyangfl@gmail.com Signed-off-by: Jakub Kicinski

net: dsa: motorcomm: Move to subdirectory

2026-09-03T02:58:54+00:00

yt921x is already the longest single-file DSA driver, so it's time to split it into parts. Signed-off-by: David Yang Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20260831044818.4103477-3-mmyangfl@gmail.com Signed-off-by: Jakub Kicinski

net: dsa: yt921x: Check lock status with lockdep_assert_held_once()

2026-09-03T02:58:53+00:00

mutex_is_locked() was used to check the lock status. While it should not be triggered unless a bug is introduced, it does not warn on lock being held by another thread. Fix it with lockdep helpers. Fixes: 186623f4aa72 ("net: dsa: yt921x: Add support for Motorcomm YT921x") Signed-off-by: David Yang Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20260831044818.4103477-2-mmyangfl@gmail.com Signed-off-by: Jakub Kicinski

net: dsa: mt7530: implement port_change_conduit op

2026-09-03T02:10:06+00:00

Allow changing the CPU port affinity of user ports at runtime via the IFLA_DSA_CONDUIT netlink attribute. This updates the port matrix to forward to the new CPU port instead of the old one. Limit the operation to MT7531. There, trapped link-local frames follow the per-port affinity, as the MT7531_CPU_PMAP destination mask is further restricted by the port matrix. A conduit change is hence fully honoured by the hardware, for regular traffic as well as for trapped frames. The MT7530 switch, including the variant embedded in the MT7621 SoC, instead traps frames to the single CPU port set in the CPU_PORT field of the MFC register, regardless of the affinity of the inbound user port. With user ports affine to different CPU ports there is no correct value for that field, so per-port CPU affinity cannot be fully implemented for trapped frames. Routing a WAN port via the second SoC GMAC is conventionally covered by the PHY muxing feature on these switches, which bypasses the switch fabric and does not involve a CPU port at all. The switches on the MT7988, EN7581, AN7583 and EN7528 SoCs only have a single CPU port, leaving no other conduit to change to. As the op lives in the shared mt7530_switch_ops, populate the extack when rejecting the unsupported variants instead of returning a bare -EOPNOTSUPP. Also reject a conduit that belongs to a different switch in the tree, whose port index has no meaning in the local port matrix. Signed-off-by: Daniel Golle Acked-by: Chester A. Unal Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/cfc6a478a384a7f309f08d9f76be8fd029c1f202.1788190568.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski

net: dsa: mt7530: implement port_fast_age

2026-09-03T02:10:06+00:00

Implement the .port_fast_age DSA operation by flushing all non-static (dynamically learned) MAC address entries from the address table. The switch does not offer a combined "non-static AND per-port" match mode, so the flush is global and the port argument is not used. Unlike b53 and realtek, which flush the dynamic entries of the affected port only, an STP topology change on one port therefore also flushes the dynamically learned entries of the other ports; they are quickly relearned. Access the address table control register under priv->reg_mutex, as done by all other ATC users (FDB and MDB add/del/dump), to serialise the write-then-poll command sequence, and log a message should the command fail. Signed-off-by: Daniel Golle Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/60f83ade0e371dfe2d974d5d192056935c3ab5c9.1788190568.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski

net: dsa: mt7530: convert to use field accessor macros

2026-09-03T02:10:06+00:00

Use FIELD_GET and FIELD_PREP instead of open-coding register fields. Replace 0x1f constant with (PHY_MAX_ADDR - 1). Some field macros (ATC_HASH, VTCR_VID and STAG_VPID) were previously defined as object-like macros referencing an undeclared 'x' and were therefore unusable; convert them into proper FIELD_PREP() accessors. The masks are equivalent to the open-coded values they replace, with one exception: mt7530_vlan_cmd() now masks vid through VTCR_VID(vid) instead of OR-ing it in raw, guarding the adjacent VTCR_FUNC field against any bits set above the 12-bit VID width. VLAN IDs are always in 0-4094, so this is a no-op in practice; it is a real hardening rather than a strictly equivalent mechanical substitution. Signed-off-by: Daniel Golle Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/bd4eaf10fb6aa7a1bd056514b61b4481175708a5.1788190568.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski

net: dsa: mt7530: drop the dummy poll machinery

2026-09-03T02:10:06+00:00

Convert the two remaining switch reset polls to regmap_read_poll_timeout() and remove struct mt7530_dummy_poll, INIT_MT7530_DUMMY_POLL() and the mt7530_mii_poll() helper. As with the command polls addressed by the preceding fixes, a failed read now terminates the poll with an error instead of being fed to the loop condition as a zero value. Generated using the following semantic patch, plus removal of the two now-unused definitions from mt7530.h: // Poll the register through regmap so a failed read ends the poll. @@ expression priv, reg; identifier p, ret, val; @@ - INIT_MT7530_DUMMY_POLL(&p, priv, reg); - ret = readx_poll_timeout(mt7530_mii_poll, &p, val, val != 0, - 20, 1000000); + ret = regmap_read_poll_timeout(priv->regmap, reg, val, val != 0, + 20, 1000000); @@ identifier p; @@ - struct mt7530_dummy_poll p; ... when != p @@ identifier p; type T; @@ -static T mt7530_mii_poll(struct mt7530_dummy_poll *p) { ... } Signed-off-by: Daniel Golle Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/609dcd8dd0c335e36bbc8df41413da58af1dbfb4.1788190568.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski

net: dsa: mt7530: replace mt7530_read with regmap_read

2026-09-03T02:10:06+00:00

Replace all mt7530_read() calls with direct regmap_read() calls and remove the wrapper function. The wrapper returned 0 on failed reads; preserve that behaviour by zeroing the read-back buffer at the start of mt7530_regmap_read(), so callers that do not check the return code keep seeing a deterministic value. The WARN_ON_ONCE error logging is dropped. Most callsites follow the val = mt7530_read(priv, reg) pattern and are converted mechanically using the following semantic patch: @@ expression priv, reg; identifier val; @@ -val = mt7530_read(priv, reg); +regmap_read(priv->regmap, reg, &val); Remaining inline uses are converted by hand. Signed-off-by: Daniel Golle Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/a1dc338ae368554ec6faa974df7ba6f3660ae12b.1788190568.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski

net: dsa: mt7530: replace mt7530_rmw/set/clear with regmap API

2026-09-03T02:10:06+00:00

Replace all mt7530_rmw() calls with regmap_update_bits(), mt7530_set() with regmap_set_bits(), and mt7530_clear() with regmap_clear_bits(). Remove the wrapper function definitions. Generated using the following semantic patch: @@ expression priv, reg, mask, set; @@ -mt7530_rmw(priv, reg, mask, set) +regmap_update_bits(priv->regmap, reg, mask, set) @@ expression priv, reg, val; @@ -mt7530_set(priv, reg, val) +regmap_set_bits(priv->regmap, reg, val) @@ expression priv, reg, val; @@ -mt7530_clear(priv, reg, val) +regmap_clear_bits(priv->regmap, reg, val) Six regmap_update_bits() call sites with multi-line masks are rewrapped by hand afterwards, as spatch joins their argument lines. Signed-off-by: Daniel Golle Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/182be19b483e77eb1d4ed835813359a27acd8e9c.1788190568.git.daniel@makrotopia.org Signed-off-by: Jakub Kicinski