From: sashiko-bot@kernel.org
To: "Florian Fainelli" <florian.fainelli@broadcom.com>
Cc: "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS"
<devicetree@vger.kernel.org>,
robh@kernel.org, conor+dt@kernel.org
Subject: Re: [PATCH net-next 5/9] net: dsa: add Broadcom FA metadata and IPv4 exception paths
Date: Tue, 17 Nov 2026 22:33:59 +0000 [thread overview]
Message-ID: <sashiko-outbox-219753@kernel.org> (raw)
In-Reply-To: <20261116222519.1416501-6-florian.fainelli@broadcom.com>
Thank you for your contribution! Sashiko AI review found 7 potential
issues to consider:
- [Medium] net: dsa: b53: broken guard error handling in
b53_fa_netdev_event()
- [Medium] net: dsa: b53: unrecorded standalone port on b53_fa_port()
failure
- [Medium] net: dsa: b53: missing CPU port 5 in non-default VLANs in
b53_vlan_add()
- [Medium] net: dsa: b53: active fa_netdev_nb during teardown and
b53_switch_shutdown()
- [Medium] net: dsa: tag_brcm: brcm_tag_rcv_ll() reads prepended tag as
MAC DA
- [Low] net: dsa: b53: duplicate CFP rule programming in b53_fa_setup()
- [Low] net: dsa: b53: unlocked fa_guards_ready access in
b53_switch_register()
--
commit 5efa0c79a851eb8221145c59826df5a46728bb31
Author: Florian Fainelli <florian.fainelli@broadcom.com>
net: dsa: add Broadcom FA metadata and IPv4 exception paths
This commit adds a brcm-fa DSA tag driver for the Broadcom Northstar flow
accelerator on port 8 and configures B53 CFP rules to redirect IPv4
exception traffic and router-MAC guard misses to auxiliary CPU port 5.
> diff --git a/drivers/net/dsa/b53/b53_common.c b/drivers/net/dsa/b53/b53_common.c
> index b7aff5e238e85..3e8c9679836d3 100644
> --- a/drivers/net/dsa/b53/b53_common.c
> +++ b/drivers/net/dsa/b53/b53_common.c
[ ... ]
> +static int b53_fa_rule(struct b53_device *dev, unsigned int index,
> + const u32 data[8], const u32 mask[8], bool trap)
> +{
> + unsigned int i;
> + int err;
> +
> + for (i = 0; i < 8; i++) {
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_DATA(i), data[i]);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_MASK(i), mask[i]);
> + if (err)
> + return err;
> + }
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_ACTION(0),
> + trap ? CFP_ACT0_FWD_MAP_REPLACE | CFP_ACT0_DST_PORT5 |
> + CFP_ACT0_REASON_TERMINATION : 0);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_ACTION(1),
> + trap ? CFP_ACT1_CLASS_ID(1) | CFP_ACT1_FWD_MAP_REPLACE |
> + CFP_ACT1_DST_PORT5 : 0);
> + if (!err)
> + err = b53_fa_cfp_write(dev, index, CFP_ACC_ACTION);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_METER, 0);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_METER + 4, 0);
> + if (!err)
> + err = b53_fa_cfp_write(dev, index, CFP_ACC_METER);
> + return err ?: b53_fa_cfp_write(dev, index, CFP_ACC_TCAM);
> +}
[ ... ]
> +static int b53_fa_setup(struct b53_device *dev)
> +{
[ ... ]
> + for (rule = 0; rule < 3; rule++) {
> + memset(data, 0, sizeof(data));
> + memset(mask, 0, sizeof(mask));
> + data[0] = CFP_TCAM0_VALID | CFP_TCAM0_SLICE1 | CFP_TCAM0_UDF0_FLAG(rule);
> + /* Slice 1; FIN, SYN or RST. */
> + mask[0] = CFP_TCAM0_VALID | CFP_TCAM0_SLICE_MASK | CFP_TCAM0_UDF0_FLAG(rule);
> + data[5] = CFP_TCAM5_UDF_VALID(0); /* UDF A0 valid. */
> + mask[5] = CFP_TCAM5_UDF_VALID(0);
> + data[6] = FIELD_PREP(CFP_TCAM6_IP_PROTOCOL, IPPROTO_TCP);
> + mask[6] = CFP_TCAM6_L3_FRAMING | CFP_TCAM6_IP_PROTOCOL | CFP_TCAM6_IP_FRAGMENT;
> + for (i = 0; i < ARRAY_SIZE(data); i++) {
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_DATA(i), data[i]);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_MASK(i), mask[i]);
> + if (err)
> + return err;
> + }
> + /* Both meter outcomes redirect to IMP5, with class 1 and a
> + * termination reason. Preserve VLAN, EAP and STP filtering.
> + */
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_ACTION(0),
> + CFP_ACT0_FWD_MAP_REPLACE | CFP_ACT0_DST_PORT5 |
> + CFP_ACT0_REASON_TERMINATION);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_ACTION(1),
> + CFP_ACT1_CLASS_ID(1) | CFP_ACT1_FWD_MAP_REPLACE |
> + CFP_ACT1_DST_PORT5);
> + if (!err)
> + err = b53_fa_cfp_write(dev, rule, CFP_ACC_ACTION);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_METER, 0);
> + if (!err)
> + err = b53_write32(dev, B53_CFP_TCAM_PAGE, B53_CFP_METER + 4, 0);
> + if (!err)
> + err = b53_fa_cfp_write(dev, rule, CFP_ACC_METER);
> + if (!err)
> + err = b53_fa_cfp_write(dev, rule, CFP_ACC_TCAM);
> + if (err)
> + return err;
> + }
[Severity: Low]
This isn't a bug, but this loop open-codes b53_fa_rule(dev, rule, data,
mask, true).
[ ... ]
> +static int b53_fa_port(struct b53_device *dev, int port, bool standalone)
> +{
> + struct dsa_port *control = dsa_to_port(dev->ds, 5);
> + bool ready;
> + u16 ports;
> + int err;
> +
> + if (!dev->fa_cfp_enabled)
> + return 0;
> + ports = dev->fa_cfp_ports & ~BIT(port);
> + if (standalone)
> + ports |= BIT(port);
> + ready = b53_fa_ready(dev, dsa_port_conduit_is_operational(control));
> + err = b53_write16(dev, B53_CFP_PAGE, B53_CFP_CTRL,
> + ready ? ports : 0);
> + if (!err)
> + dev->fa_cfp_ports = ports;
> + return err;
> +}
[ ... ]
> +static int b53_fa_netdev_event(struct notifier_block *nb, unsigned long event, void *ptr)
> +{
> + struct b53_device *dev = container_of(nb, struct b53_device, fa_netdev_nb);
> + struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
> + struct dsa_port *dp, *control, *data_cpu;
> + u32 data[8] = {}, mask[8] = {};
> + unsigned int i;
> + int err;
> +
> + if (dev->fa_stopping || !dsa_user_dev_check(netdev) ||
> + (event != NETDEV_REGISTER && event != NETDEV_UP &&
> + event != NETDEV_PRE_CHANGEADDR && event != NETDEV_CHANGEADDR))
> + return NOTIFY_DONE;
> + dp = dsa_port_from_netdev(netdev);
> + if (dp->ds != dev->ds || dp->index >= 8)
> + return NOTIFY_DONE;
> + control = dsa_to_port(dev->ds, 5);
> + data_cpu = dsa_to_port(dev->ds, 8);
> + brcm_fa_control_set_ready(data_cpu->conduit, false);
> + if (event != NETDEV_PRE_CHANGEADDR) {
> + data[0] = CFP_TCAM0_VALID | CFP_TCAM0_SLICE1;
> + mask[0] = CFP_TCAM0_VALID | CFP_TCAM0_SLICE_MASK;
> + mask[6] = CFP_TCAM6_L3_FRAMING;
> + data[7] = BIT(dp->index);
> + mask[7] = CFP_TCAM7_SOURCE_PORTS;
> + for (i = 0; i < 3; i++)
> + b53_fa_udf(data, mask, 3 + i, get_unaligned_be16(netdev->dev_addr + 2 * i),
> + 0xffff);
> + }
> + err = b53_fa_rule(dev, B53_FA_ALLOW_BASE + dp->index, data, mask, false);
> + if (err) {
> + dev->fa_guards_failed = true;
> + dev_err(dev->dev, "failed to update FA ingress guard: %d\n", err);
> + } else if (event != NETDEV_PRE_CHANGEADDR) {
> + b53_fa_conduit_state_change(dev->ds, control->conduit,
> + dsa_port_conduit_is_operational(control));
> + }
> + return NOTIFY_DONE;
> +}
[Severity: Medium]
What restores the port's allow rule at B53_FA_ALLOW_BASE + dp->index if
netif_set_mac_address() fails after NETDEV_PRE_CHANGEADDR?
netif_set_mac_address() emits NETDEV_PRE_CHANGEADDR before calling
ops->ndo_set_mac_address(). If a subsequent notifier or
dsa_user_set_mac_address() rejects the address (for example when
!is_valid_ether_addr()), NETDEV_CHANGEADDR is never sent. The port's
hardware allow rule remains zeroed out and all IPv4 traffic on that port
hits B53_FA_DEFAULT_RULE and traps to CPU port 5, even after another port
event re-enables FA via b53_fa_conduit_state_change().
Also, when b53_fa_rule() fails here, dev->fa_guards_failed is set to true,
but skipping b53_fa_conduit_state_change() leaves B53_CFP_CTRL enabled in
hardware.
[ ... ]
> @@ -2439,6 +2775,10 @@ void b53_br_leave(struct dsa_switch *ds, int port, struct dsa_bridge bridge)
> vl->members |= BIT(port);
> b53_set_vlan_entry(dev, pvid, vl);
> }
> + if (b53_fa_port(dev, port, true)) {
> + dev_err(ds->dev, "failed to restore FA classification on port %d\n", port);
> + b53_disable_port(ds, port);
> + }
> }
[Severity: Medium]
If b53_fa_port(dev, port, true) fails in b53_br_leave(), can a subsequent
ndo_open re-enable the standalone port without CFP classification?
When b53_write16() fails in b53_fa_port(), BIT(port) is not recorded in
dev->fa_cfp_ports and dev->fa_guards_failed is not set. Bringing the port
up later via dsa_user_open() -> b53_enable_port() re-enables Rx/Tx without
setting BIT(5) in B53_PVLAN_PORT_MASK(port), and the following NETDEV_UP
event in b53_fa_netdev_event() writes dev->fa_cfp_ports (still missing
BIT(port)) to B53_CFP_CTRL and re-enables FA admission.
[ ... ]
> +static struct dsa_port *b53_preferred_cpu_port(struct dsa_switch *ds)
> +{
> + struct b53_device *dev = ds->priv;
> +
> + if (is5301x(dev) && dsa_is_cpu_port(ds, B53_CPU_PORT) && b53_uses_fa_tag(ds, B53_CPU_PORT))
> + return dsa_to_port(ds, B53_CPU_PORT);
> + return NULL;
> +}
[Severity: Medium]
When dev->vlan_filtering is enabled, will hardware VLAN filtering drop
VLAN-tagged IPv4 frames that CFP redirects to CPU port 5?
CFP trap rules redirect frames to port 5 without bypassing VLAN filtering,
and on B53 ds->vlan_filtering_is_global is true, so enabling VLAN filtering
on any bridge enforces VLAN table membership across all ports. Because
b53_preferred_cpu_port() assigns port 8 as dp->cpu_dp for all user ports,
DSA host VLAN programming only calls b53_vlan_add() for the user port and
port 8, leaving BIT(5) unset in dev->vlans[vid].members for any non-default
VID.
[ ... ]
> @@ -3385,7 +3773,32 @@ int b53_switch_register(struct b53_device *dev)
> dev_info(dev->dev, "found switch: %s, rev %i\n",
> dev->name, dev->core_rev);
>
> - return dsa_register_switch(dev->ds);
> + ret = dsa_register_switch(dev->ds);
> + if (ret || !dev->fa_cfp_enabled)
> + return ret;
> + /* Registration replays existing user devices, after their MACs and
> + * DSA private data are initialized. Keep lookup gated until it finishes.
> + */
> + dev->fa_netdev_nb.notifier_call = b53_fa_netdev_event;
> + ret = register_netdevice_notifier(&dev->fa_netdev_nb);
> + if (ret)
> + goto unregister_switch;
> + ret = devm_add_action_or_reset(dev->dev, b53_fa_unregister_notifier, dev);
> + if (ret)
> + goto unregister_switch;
[Severity: Medium]
Can b53_fa_netdev_event() run while the switch is being torn down or after
b53_switch_shutdown()?
Deferring b53_fa_unregister_notifier() to devres keeps dev->fa_netdev_nb
registered with dev->fa_stopping == false throughout
dsa_tree_teardown_conduit() and dsa_tree_teardown_ports(), because
b53_teardown() only runs later in dsa_tree_teardown_switches(). During
system shutdown, b53_switch_shutdown() -> dsa_switch_shutdown() drops the
netdev_put() references on each CPU port's conduit without calling
b53_teardown() or running devres actions, so any later user port netdevice
event will dereference the unreferenced control->conduit and
data_cpu->conduit pointers.
> + dev->fa_guards_ready = true;
> + if (!dev->fa_guards_failed) {
> + struct dsa_port *control = dsa_to_port(dev->ds, 5);
> +
> + rtnl_lock();
> + b53_fa_conduit_state_change(dev->ds, control->conduit,
> + dsa_port_conduit_is_operational(control));
> + rtnl_unlock();
> + }
[Severity: Low]
Should dev->fa_guards_ready and dev->fa_guards_failed be accessed under
rtnl_lock() here? register_netdevice_notifier() has already dropped
rtnl_lock(), so a concurrent netdevice event can run b53_fa_netdev_event()
or b53_fa_port() on another CPU while these booleans are updated and read.
> diff --git a/net/dsa/tag_brcm.c b/net/dsa/tag_brcm.c
> index b7c49822ca888..8d13a3c7de5e3 100644
> --- a/net/dsa/tag_brcm.c
> +++ b/net/dsa/tag_brcm.c
[ ... ]
> @@ -146,9 +150,10 @@ static struct sk_buff *brcm_tag_xmit_ll(struct sk_buff *skb,
> */
> static struct sk_buff *brcm_tag_rcv_ll(struct sk_buff *skb,
> struct net_device *dev,
> - unsigned int offset)
> + unsigned int offset, bool fa)
> {
[ ... ]
> @@ -407,7 +416,100 @@ DSA_TAG_DRIVER(brcm_prepend_netdev_ops);
[ ... ]
> +static struct sk_buff *brcm_fa_tag_rcv(struct sk_buff *skb, struct net_device *dev)
> +{
> + struct dsa_port *cpu = dev->dsa_ptr;
> + struct brcm_fa_control *control;
> + struct dsa_port *data_cpu;
> + unsigned int port;
> + u8 reason;
> +
> + if (cpu->index != 5)
> + return brcm_tag_rcv_ll(skb, dev, ETH_HLEN, true);
[Severity: Medium]
On the cpu->index != 5 path, does brcm_tag_rcv_ll() inspect the 4-byte
prepended FA tag instead of the destination MAC when checking
is_link_local_ether_addr(eth_hdr(skb)->h_dest)?
net/dsa/tag_brcm.c:brcm_tag_rcv_ll() {
...
/* Remove Broadcom tag and update checksum */
skb_pull_rcsum(skb, BRCM_TAG_LEN);
if (likely(!is_link_local_ether_addr(eth_hdr(skb)->h_dest)))
dsa_default_offload_fwd_mark(skb);
...
}
eth_type_trans() on the conduit sets skb->mac_header to the start of the
prepended tag, and skb_pull_rcsum(skb, BRCM_TAG_LEN) only advances
skb->data; skb->mac_header is not updated until dsa_switch_rcv() runs after
cpu_dp->rcv() returns. Because byte 0 of the prepended tag does not have
bit 0 set, is_link_local_ether_addr() returns false and link-local control
frames (01:80:c2:00:00:xx) on bridged ports get skb->offload_fwd_mark set,
preventing the software bridge from forwarding them when STP is disabled or
group_fwd_mask is configured.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261116222519.1416501-1-florian.fainelli@broadcom.com?part=5
next prev parent reply other threads:[~2026-11-17 22:33 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-16 22:25 [PATCH net-next 0/9] net: bgmac: add Northstar IPv4 TCP/UDP NAT flowtable offload Florian Fainelli
2026-11-16 22:25 ` [PATCH net-next 1/9] net: ethernet: bgmac: validate RX length against the DMA packet capacity Florian Fainelli
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-17 23:43 ` Florian Fainelli
2026-11-16 22:25 ` [PATCH net-next 2/9] netfilter: flowtable: reject non-host IPv4 packets before lookup Florian Fainelli
2026-11-17 22:33 ` sashiko-bot
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-17 23:51 ` Florian Fainelli
2026-11-16 22:25 ` [PATCH net-next 3/9] netfilter: flowtable: retain offload ownership until deletion completes Florian Fainelli
2026-11-17 22:33 ` sashiko-bot
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-17 23:57 ` Florian Fainelli
2026-11-16 22:25 ` [PATCH net-next 4/9] dt-bindings: net: describe the Northstar flow accelerator Florian Fainelli
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-16 22:25 ` [PATCH net-next 5/9] net: dsa: add Broadcom FA metadata and IPv4 exception paths Florian Fainelli
2026-11-17 22:33 ` sashiko-bot [this message]
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-16 22:25 ` [PATCH net-next 6/9] net: flow_offload: allow drivers to validate and invalidate cached paths Florian Fainelli
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-16 22:25 ` [PATCH net-next 7/9] net: bgmac: add Northstar IPv4 TCP/UDP NAT flowtable offload Florian Fainelli
2026-11-17 22:33 ` sashiko-bot
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-16 22:25 ` [PATCH net-next 8/9] net: add KUnit coverage for FA and flowtable path validation Florian Fainelli
2026-11-17 22:34 ` sashiko-bot
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-16 22:25 ` [PATCH net-next 9/9] docs: net: document Northstar FA setup and NAT qualification Florian Fainelli
2026-11-17 23:29 ` netdev-bot+sashiko
2026-11-16 22:38 ` [PATCH net-next 0/9] net: bgmac: add Northstar IPv4 TCP/UDP NAT flowtable offload netdev-bot+sinfo
2026-11-16 22:45 ` Florian Fainelli
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-219753@kernel.org \
--to=sashiko-bot@kernel.org \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=florian.fainelli@broadcom.com \
--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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.