From: sashiko-bot@kernel.org
To: "Caleb James DeLisle" <cjd@cjdns.fr>
Cc: devicetree@vger.kernel.org, robh@kernel.org, conor+dt@kernel.org
Subject: Re: [net-next v5 3/7] net: dsa: mediatek: add support for passthrough mode
Date: Tue, 17 Nov 2026 18:35:06 +0000 [thread overview]
Message-ID: <sashiko-outbox-219352@kernel.org> (raw)
In-Reply-To: <20261116183332.168858-4-cjd@cjdns.fr>
Thank you for your contribution! Sashiko AI review found 5 potential
issues to consider:
- [Medium] net: dsa: mt7530: wrong upstream port in
mt753x_conduit_state_change()
- [Medium] net: dsa: mt7530: broken CPU port fallback in
mt753x_upstream_port_index()
- [Medium] net: dsa: mt7530: NULL dp->user dereference in
mt7531_mac_config() on DSA port
- [Medium] net: dsa: mt7530: unchecked MT753X_MTRAP write in
mt753x_cpu_port_enable()
- [Low] net: dsa: mt7530: dead no-op callbacks in
mt7530_passthrough_switch_ops
--
commit c217d406c794dac728f46979f906b1ad1d787a4d
Author: Caleb James DeLisle <cjd@cjdns.fr>
net: dsa: mediatek: add support for passthrough mode
This commit adds passthrough mode support to the mt7530 DSA driver so an
upstream MediaTek switch can forward traffic unaltered to and from a
downstream MediaTek switch.
> diff --git a/drivers/net/dsa/mt7530.c b/drivers/net/dsa/mt7530.c
> index 3bfe29b3b93e2..09b7e9d2d0253 100644
> --- a/drivers/net/dsa/mt7530.c
> +++ b/drivers/net/dsa/mt7530.c
[ ... ]
> @@ -1286,13 +1297,81 @@ mt753x_trap_frames(struct mt7530_priv *priv)
> +/* If this switch is downstream of another switch that is in passthrough mode,
> + * the "CPU" port is actually a DSA port.
> + */
> static void
> mt753x_cpu_port_enable(struct dsa_switch *ds, int port)
> {
> struct mt7530_priv *priv = ds->priv;
>
> - /* Enable Mediatek header mode on the cpu port */
> - regmap_write(priv->regmap, MT7530_PVC_P(port), PORT_SPEC_TAG);
> + if (priv->is_passthrough) {
> + /* Disable parsing of the DSA tag, it will be forwarded blindly
> + * to the downstream switch.
> + */
> + regmap_write(priv->regmap, MT7530_PVC_P(port),
> + VLAN_ATTR(MT7530_VLAN_TRANSPARENT) |
> + PVC_EG_TAG(MT7530_VLAN_EG_DISABLED));
> +
> + /* The port is not configured to parse DSA tags, so they are
> + * liable to be confused for length fields, so length check is
> + * disabled.
> + */
> + regmap_clear_bits(priv->regmap, MT753X_AGC, AGC_L2LEN_CHK);
> +
> + /* In passthrough mode, MTU is only enforced downstream */
> + mt7530_port_change_mtu(ds, port, MT7530_MAX_MTU);
> +
> + /* Loop detection has no value in passthrough mode */
> + regmap_set_bits(priv->regmap, MT753X_MTRAP,
> + MT7530_LOOP_DET_DISABLE);
[Severity: Medium]
Should this MT753X_MTRAP write be guarded by a priv->id check?
mt753x_cpu_port_enable() runs from both mt7530_setup() and
mt7531_setup_common(), and MT753X_MTRAP (0x7804) has a different bit layout
on MT7531-family switches (where other MT7530_* trap accesses in
mt7530_port_enable() and mt7530_port_disable() check priv->id first). Also,
on ID_EN7528 (an MT7530 core initialized via mt7988_setup()),
MT7530_CHG_TRAP is never set before modifying MT753X_MTRAP.
[ ... ]
> @@ -1318,12 +1397,36 @@ mt753x_cpu_port_enable(struct dsa_switch *ds, int port)
> PCR_PORT_VLAN_MASK, MT7530_PORT_FALLBACK_MODE);
> }
>
> +/**
> + * mt753x_upstream_port_index - get the index of the upstream port.
> + *
> + * If the CPU port is local then this is the upstream port, but in the case of
> + * hierarchy the CPU port may be on another switch and the upstream port is
> + * then the DSA port pointing toward that CPU port. Note that
> + * dsa_switch_upstream_port() is not reliable because the conduit for a given
> + * port can be customized.
> + */
> +static int
> +mt753x_upstream_port_index(struct dsa_port *dp)
> +{
> + struct dsa_port *cpu_dp = dp->cpu_dp;
> +
> + /* Early port_enable on CPU port happens before cpu_dp is set. */
> + if (!cpu_dp)
> + return dsa_switch_upstream_port(dp->ds);
> +
> + return dsa_towards_port(dp->ds, cpu_dp->ds->index, cpu_dp->index);
> +}
[Severity: Medium]
Does mt753x_upstream_port_index() return the wrong port when called on a
non-default CPU port?
In DSA, dp->cpu_dp is only populated on user and DSA ports and remains NULL
on CPU ports. On a switch with multiple CPU ports (such as MT7531 with
ports 5 and 6), calling mt753x_upstream_port_index() on CPU port 5 hits the
!cpu_dp branch and returns dsa_switch_upstream_port(dp->ds) (port 6). In
mt7530_hw_vlan_add(), the check:
if (dp->index == mt753x_upstream_port_index(dp))
val = MT7530_VLAN_EGRESS_STACK;
then evaluates to false for CPU port 5, configuring it as untagged or
tagged instead of MT7530_VLAN_EGRESS_STACK.
dsa_upstream_port(ds, port) and dsa_is_upstream_port(ds, port) in
include/net/dsa.h already handle both per-port dp->cpu_dp routing and
returning port directly when dp->cpu_dp is NULL, so using those helpers
avoids this regression. Also, the /** kernel-doc comment on
mt753x_upstream_port_index() omits @dp, which triggers a W=1 kernel-doc
warning.
[ ... ]
> @@ -2635,7 +2818,7 @@ mt7531_setup_common(struct dsa_switch *ds)
> regmap_set_bits(priv->regmap, MT7531_DBG_CNT(i),
> MT7531_DIS_CLR);
>
> - if (dsa_is_cpu_port(ds, i)) {
> + if (dsa_is_upstream_port(ds, i)) {
> mt753x_cpu_port_enable(ds, i);
> } else {
> mt7530_port_disable(ds, i);
[Severity: Medium]
Can enabling a DSA cascade port on MT7531 cause a NULL pointer dereference
in mt7531_mac_config()?
When port 5 on an MT7531 switch is configured in RGMII mode as an upstream
or downstream DSA link port (DSA_PORT_TYPE_DSA), dp->user is NULL, but
mt7531_mac_config() unconditionally dereferences it during phylink startup:
drivers/net/dsa/mt7530.c:mt7531_mac_config() {
...
if (phy_interface_mode_is_rgmii(interface)) {
dp = dsa_to_port(ds, port);
phydev = dp->user->phydev;
mt7531_rgmii_setup(priv, interface, phydev);
}
...
}
[ ... ]
> @@ -3402,6 +3600,96 @@ static int mt7988_setup(struct dsa_switch *ds)
> +/* No manipulation of forwarding rules allowed in passthrough mode */
> +static const struct dsa_switch_ops mt7530_passthrough_switch_ops = {
> + .get_tag_protocol = mtk_get_tag_protocol,
> + .setup = mt753x_setup,
> + .teardown = mt753x_teardown,
> + .preferred_default_local_cpu_port = mt753x_preferred_default_local_cpu_port,
> + .port_change_conduit = mt753x_passthrough_port_change_conduit,
> + .get_strings = mt7530_get_strings,
> + .get_ethtool_stats = mt7530_get_ethtool_stats,
> + .get_sset_count = mt7530_get_sset_count,
> + .get_eth_mac_stats = mt7530_get_eth_mac_stats,
> + .get_rmon_stats = mt7530_get_rmon_stats,
> + .get_eth_ctrl_stats = mt7530_get_eth_ctrl_stats,
> + .get_stats64 = mt7530_get_stats64,
> + .set_ageing_time = mt7530_passthrough_set_ageing_time,
> + .port_enable = mt7530_port_enable,
> + .port_disable = mt7530_port_disable,
> + .port_change_mtu = mt7530_passthrough_port_change_mtu,
> + .port_max_mtu = mt7530_port_max_mtu,
> + .port_stp_state_set = mt7530_passthrough_stp_state_set,
> + .port_pre_bridge_flags = mt7530_port_pre_bridge_flags,
> + .port_bridge_flags = mt7530_passthrough_port_bridge_flags,
> + .port_bridge_join = mt7530_passthrough_port_bridge_join,
> + .port_bridge_leave = mt7530_passthrough_port_bridge_leave,
> + .port_fast_age = mt7530_passthrough_port_fast_age,
> + .port_fdb_add = mt7530_passthrough_port_fdb_add,
> + .port_fdb_del = mt7530_passthrough_port_fdb_del,
> + .port_fdb_dump = mt7530_port_fdb_dump,
> + .port_mdb_add = mt7530_passthrough_port_mdb_add,
> + .port_mdb_del = mt7530_passthrough_port_mdb_del,
> + .port_vlan_filtering = mt7530_passthrough_port_vlan_filtering,
> + .port_vlan_add = mt7530_passthrough_port_vlan_add,
> + .port_vlan_del = mt7530_passthrough_port_vlan_del,
> + .port_mirror_add = mt753x_passthrough_port_mirror_add,
> + .port_mirror_del = mt753x_passthrough_port_mirror_del,
> + .phylink_get_caps = mt753x_phylink_get_caps,
> + .support_eee = dsa_supports_eee,
> + .set_mac_eee = mt753x_set_mac_eee,
[Severity: Low]
This isn't a bug, but most of the new no-op callbacks in
mt7530_passthrough_switch_ops appear unnecessary. Because
mt753x_check_passthrough() requires the switch to have only one CPU port
and one DSA link port (and no user ports), only the tree-wide notifier
callbacks (.port_change_mtu, .port_fdb_add, .port_fdb_del, .port_mdb_add,
.port_mdb_del, .port_vlan_add, and .port_vlan_del) need no-op stubs to
avoid returning -EOPNOTSUPP.
.set_ageing_time, .port_stp_state_set, and .port_fast_age already default
to no-ops in the DSA core when left NULL, while .port_bridge_join,
.port_bridge_leave, .port_bridge_flags, .port_vlan_filtering,
.port_mirror_add, .port_mirror_del, .port_change_conduit, and
.port_setup_tc (along with user-port-only callbacks like
.port_pre_bridge_flags, .port_fdb_dump, .port_max_mtu, .support_eee,
.set_mac_eee, .port_hsr_join, and .port_hsr_leave) are never called on a
switch with no user ports.
> + .conduit_state_change = mt753x_conduit_state_change,
[Severity: Medium]
Does mt753x_conduit_state_change() program the wrong CPU port on a
downstream switch?
DSA broadcasts DSA_NOTIFIER_CONDUIT_STATE_CHANGE to every switch in the
tree, with cpu_dp = conduit->dsa_ptr pointing to the CPU port on the
upstream switch:
drivers/net/dsa/mt7530.c:mt753x_conduit_state_change() {
...
mask = BIT(cpu_dp->index);
if (operational)
priv->active_cpu_ports |= mask;
else
priv->active_cpu_ports &= ~mask;
if (priv->active_cpu_ports) {
val = MT7530_CPU_EN |
MT7530_CPU_PORT(__ffs(priv->active_cpu_ports));
}
regmap_update_bits(priv->regmap, MT753X_MFC,
MT7530_CPU_EN | MT7530_CPU_PORT_MASK, val);
...
}
When a downstream switch's local upstream DSA port index differs from the
upstream switch's CPU port index (for example, upstream CPU port 6 wired to
downstream port 5), MT7530_CPU_PORT is set to cpu_dp->index instead of
dsa_towards_port(ds, cpu_dp->ds->index, cpu_dp->index), sending trapped
control frames (BPDUs, PAE, LLDP, LACP) to the wrong port on the downstream
switch. In addition, wiring mt753x_conduit_state_change() in
mt7530_passthrough_switch_ops sets MT7530_CPU_EN on the passthrough switch
even though trapping is disabled in passthrough mode.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261116183332.168858-1-cjd@cjdns.fr?part=3
next prev parent reply other threads:[~2026-11-17 18:35 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-16 18:33 [net-next v5 0/7] net: dsa: mt7530: support EcoNet EN751221 Caleb James DeLisle
2026-11-16 18:33 ` [net-next v5 1/7] net: dsa: mt7530: get ctrl phy addr using a function Caleb James DeLisle
2026-11-16 18:33 ` [net-next v5 2/7] dt-bindings: net: dsa: mediatek,mt7530: describe 'passthrough' hierarchy Caleb James DeLisle
2026-11-16 18:33 ` [net-next v5 3/7] net: dsa: mediatek: add support for passthrough mode Caleb James DeLisle
2026-11-17 0:30 ` kernel test robot
2026-11-17 18:35 ` sashiko-bot [this message]
2026-11-16 18:33 ` [net-next v5 4/7] net: dsa: mediatek: support PLL setup on MMIO MT7530 Caleb James DeLisle
2026-11-17 18:35 ` sashiko-bot
2026-11-16 18:33 ` [net-next v5 5/7] net: dsa: mediatek: support MDIO switch downstream of MMIO switch Caleb James DeLisle
2026-11-17 18:35 ` sashiko-bot
2026-11-16 18:33 ` [net-next v5 6/7] dt-bindings: net: dsa: mediatek,mt7530: add EcoNet EN751221 switches Caleb James DeLisle
2026-11-17 18:35 ` sashiko-bot
2026-11-16 18:33 ` [net-next v5 7/7] net: dsa: mediatek: support EN751221 switch Caleb James DeLisle
2026-11-17 18:35 ` sashiko-bot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=sashiko-outbox-219352@kernel.org \
--to=sashiko-bot@kernel.org \
--cc=cjd@cjdns.fr \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=robh@kernel.org \
--cc=sashiko-reviews@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox