Re: [PATCH net-next v4 2/2] net: airoha: defer GDM3/GDM4 WAN mode and GDM2 loopback to QoS offload

[Lorenzo Bianconi <lorenzo@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 13669 bytes --]

> GDM3 and GDM4 ports require GDM2 loopback to be enabled for hardware
> QoS offload to function. Without it, HTB and ETS offload on these ports
> do not work.
> Previously, GDM3/GDM4 ports were automatically configured as WAN with
> GDM2 loopback enabled during ndo_init(). Add the capability to configure
> GDM3/GDM4 as WAN/LAN on demand when QoS offload is created or destroyed.
> Hook airoha_enable_qos_for_gdm34() into TC_HTB_CREATE so that requesting
> HTB offload on a GDM3/GDM4 LAN port switches it to WAN mode and enables
> GDM2 loopback, with proper rollback on failure. Hook the counterpart
> airoha_disable_qos_for_gdm34() into TC_HTB_DESTROY to restore LAN mode
> when the offloaded qdisc is torn down.
> Since airoha_dev_set_qdma() can now be called on a running device to
> migrate between QDMA blocks, make dev->qdma an RCU pointer so the TX
> path can safely dereference it without holding RTNL.
> Hold flow_offload_mutex in airoha_dev_set_qdma() around the QDMA pointer
> update and __airoha_ppe_set_cpu_port() call, serializing against
> concurrent airoha_ppe_hw_init() in the TC_SETUP_CLSFLOWER offload path.
> Introduce airoha_qdma_deref() helper that wraps rcu_dereference_protected()
> with a lockdep condition accepting either rtnl_lock or flow_offload_mutex,
> and use it across all control-path dereferences of the RCU-protected
> dev->qdma pointer.
> Add airoha_disable_gdm2_loopback() to disable GDM2 hw loopback.
> 
> Tested-by: Madhur Agrawal <madhur.agrawal@airoha.com>
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>

Please find my comments about the following sashiko's report:
https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20260610-airoha-ethtool-priv_flags-v4-0-60e89cf28fea%40kernel.org

> ---
>  drivers/net/ethernet/airoha/airoha_eth.c  | 220 ++++++++++++++++++++++++++----
>  drivers/net/ethernet/airoha/airoha_eth.h  |  24 +++-
>  drivers/net/ethernet/airoha/airoha_ppe.c  |  18 ++-
>  drivers/net/ethernet/airoha/airoha_regs.h |   1 +
>  4 files changed, 231 insertions(+), 32 deletions(-)
> 
> diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> index aeac66df5f3b..10232470a333 100644
> --- a/drivers/net/ethernet/airoha/airoha_eth.c
> +++ b/drivers/net/ethernet/airoha/airoha_eth.c

[...]

> +static int airoha_disable_gdm2_loopback(struct airoha_gdm_dev *dev)
> +{
> +	struct airoha_gdm_port *port = dev->port;
> +	struct airoha_eth *eth = dev->eth;
> +	int i, src_port;
> +	u32 pse_port;
> +
> +	src_port = eth->soc->ops.get_sport(dev->port, dev->nbq);
> +	if (src_port < 0)
> +		return src_port;
> +
> +	airoha_fe_clear(eth,
> +			REG_SP_DFT_CPORT(src_port >> fls(SP_CPORT_DFT_MASK)),
> +			SP_CPORT_MASK(src_port & SP_CPORT_DFT_MASK));
> +
> +	airoha_fe_set(eth, REG_GDM_FWD_CFG(AIROHA_GDM2_IDX),
> +		      GDM_STRIP_CRC_MASK);
> +	airoha_set_gdm_port_fwd_cfg(eth, REG_GDM_FWD_CFG(AIROHA_GDM2_IDX),
> +				    FE_PSE_PORT_DROP);
> +	airoha_fe_clear(eth, REG_GDM_LPBK_CFG(AIROHA_GDM2_IDX),
> +			LPBK_CHAN_MASK | LPBK_MODE_MASK | LPBK_EN_MASK);
> +	pse_port = airoha_ppe_is_enabled(eth, 1) ? FE_PSE_PORT_PPE2
> +						 : FE_PSE_PORT_PPE1;
> +	airoha_set_gdm_port_fwd_cfg(eth, REG_GDM_FWD_CFG(AIROHA_GDM2_IDX),
> +				    pse_port);
> +
> +	airoha_fe_rmw(eth, REG_FE_WAN_PORT, WAN0_MASK,
> +		      FIELD_PREP(WAN0_MASK, AIROHA_GDM2_IDX));
> +
> +	for (i = 0; i < eth->soc->num_ppe; i++)
> +		airoha_ppe_clear_cpu_port(dev, i, AIROHA_GDM2_IDX);
> +
> +	/* Enable VIP and IFC for GDM2 */
> +	airoha_fe_set(eth, REG_FE_VIP_PORT_EN, BIT(AIROHA_GDM2_IDX));
> +	airoha_fe_set(eth, REG_FE_IFC_PORT_EN, BIT(AIROHA_GDM2_IDX));
> +
> +	if (port->id == AIROHA_GDM4_IDX && airoha_is_7581(eth)) {
> +		u32 mask = FC_ID_OF_SRC_PORT_MASK(dev->nbq);
> +
> +		airoha_fe_rmw(eth, REG_SRC_PORT_FC_MAP6, mask,
> +			      FC_MAP6_DEF_VALUE & mask);
> +	}
> +
> +	return 0;
> +}

- Does this disable counterpart fully undo what airoha_enable_gdm2_loopback() does? 
  I think the current implementation is correct since:
  - 0xffffffff is already the default value for REG_GDM_TXCHN_EN()
  - 0xffff is already the default value for REG_GDM_RXCHN_EN()
  - REG_GDM_LEN_CFG() will be modified by another patch (not in the series).
  - WAN1_MASK/WAN1_EN_MASK default value is 0 and the driver does not configure WAN1.
  - if the device is configured properly get_sport() callback can't fail.

> +
>  static struct airoha_gdm_dev *
>  airoha_get_wan_gdm_dev(struct airoha_eth *eth)
>  {
> @@ -2005,15 +2055,36 @@ airoha_get_wan_gdm_dev(struct airoha_eth *eth)
>  static void airoha_dev_set_qdma(struct airoha_gdm_dev *dev)
>  {
>  	struct net_device *netdev = netdev_from_priv(dev);
> +	struct airoha_qdma *cur_qdma, *qdma;
>  	struct airoha_eth *eth = dev->eth;
>  	int ppe_id;

[...]

>  }
> @@ -3027,6 +3112,89 @@ static int airoha_tc_htb_delete_leaf_queue(struct net_device *netdev,
>  	return 0;
>  }
>  
> +static int airoha_enable_qos_for_gdm34(struct net_device *netdev,
> +				       struct netlink_ext_ack *extack)
> +{
> +	struct airoha_gdm_dev *dev = netdev_priv(netdev);
> +	struct airoha_gdm_port *port = dev->port;
> +	struct airoha_eth *eth = dev->eth;
> +	int err;
> +
> +	if (port->id != AIROHA_GDM3_IDX &&
> +	    port->id != AIROHA_GDM4_IDX) {
> +		/* HW QoS is always supported by GDM1 and GDM2 */
> +		return 0;
> +	}
> +
> +	if (!airoha_is_lan_gdm_dev(dev)) /* Already enabled */
> +		return 0;
> +

- Is there a behavioural regression for GDM3/GDM4 devices that were
  auto-configured as WAN at ndo_init() time?
  I do not think there is any behavioural regression since in the current
  codebase it is not possible modify WAN/LAN configuration at runtime.
  Moreover, using tc APIs to set WAN/LAN configuration as suggested by
  Andrew, in order to configure a second device as WAN (or to set the
  current one as LAN), requires to move the current WAN device to LAN
  destroying the associated Qdisc.

> +	/* Verify the WAN device is not already configured */
> +	if (airoha_get_wan_gdm_dev(eth)) {
> +		NL_SET_ERR_MSG_MOD(extack,
> +				   "WAN device already configured");
> +		return -EBUSY;
> +	}

- The commit message says flow_offload_mutex was added to "serialize against
  concurrent airoha_ppe_hw_init() in the TC_SETUP_CLSFLOWER offload path".
  I think this is a bug and I will address the issue on the next revision.

> +
> +	dev->flags |= AIROHA_PRIV_F_WAN;
> +	airoha_dev_set_qdma(dev);
> +	err = airoha_enable_gdm2_loopback(dev);
> +	if (err)
> +		goto error_disable_wan;
> +
> +	err = airoha_set_macaddr(dev, netdev->dev_addr);
> +	if (err)
> +		goto error_disable_loopback;
> +
> +	if (netif_running(netdev)) {
> +		u32 pse_port;
> +
> +		pse_port = airoha_ppe_is_enabled(eth, 1) ? FE_PSE_PORT_PPE2
> +							 : FE_PSE_PORT_PPE1;
> +		airoha_set_gdm_port_fwd_cfg(eth, REG_GDM_FWD_CFG(port->id),
> +					    pse_port);
> +	}
> +
> +	return 0;
> +
> +error_disable_loopback:
> +	/* Restore previous LAN configuration */
> +	airoha_disable_gdm2_loopback(dev);
> +error_disable_wan:
> +	dev->flags &= ~AIROHA_PRIV_F_WAN;
> +	airoha_dev_set_qdma(dev);
> +
> +	return err;
> +}

- Is the rollback symmetric on the airoha_enable_gdm2_loopback() failure
  path?  airoha_enable_gdm2_loopback() performs many register writes
  before its only failure check (eth->soc->ops.get_sport()):
  - This has been already addressed in
    https://lore.kernel.org/netdev/20260608-airoha_enable_gdm2_loopback-minor-change-v1-1-1787a0f42b31@kernel.org/

> +
> +static void airoha_disable_qos_for_gdm34(struct net_device *netdev)
> +{
> +	struct airoha_gdm_dev *dev = netdev_priv(netdev);
> +	struct airoha_gdm_port *port = dev->port;
> +	int err;
> +
> +	if (port->id != AIROHA_GDM3_IDX &&
> +	    port->id != AIROHA_GDM4_IDX) {
> +		return;
> +	}
> +
> +	if (airoha_is_lan_gdm_dev(dev)) /* Already disabled */
> +		return;
> +
> +	err = airoha_disable_gdm2_loopback(dev);
> +	if (err)
> +		netdev_warn(netdev,
> +			    "failed disabling GDM2 loopback: %d\n", err);
> +
> +	dev->flags &= ~AIROHA_PRIV_F_WAN;
> +	airoha_dev_set_qdma(dev);
> +	airoha_set_macaddr(dev, netdev->dev_addr);
> +	if (netif_running(netdev))
> +		airoha_set_gdm_port_fwd_cfg(dev->eth,
> +					    REG_GDM_FWD_CFG(port->id),
> +					    FE_PSE_PORT_PPE1);
> +}
> +
>  static int airoha_tc_htb_destroy(struct net_device *netdev)
>  {
>  	struct airoha_gdm_dev *dev = netdev_priv(netdev);
> @@ -3035,6 +3203,8 @@ static int airoha_tc_htb_destroy(struct net_device *netdev)
>  	for_each_set_bit(q, dev->qos_sq_bmap, AIROHA_NUM_QOS_CHANNELS)
>  		airoha_tc_remove_htb_queue(netdev, q);
>  
> +	airoha_disable_qos_for_gdm34(netdev);
> +
>  	return 0;
>  }
>  
> @@ -3059,7 +3229,7 @@ static int airoha_tc_setup_qdisc_htb(struct net_device *dev,
>  {
>  	switch (opt->command) {
>  	case TC_HTB_CREATE:
> -		break;
> +		return airoha_enable_qos_for_gdm34(dev, opt->extack);

- Should ETS installed directly on a GDM3/GDM4 LAN-configured device also
  enable the loopback, or should that case be rejected with an extack
  message so the behaviour matches the description in the commit message?
  - ETS can't be used as ROOT Qdisc.

Regards,
Lorenzo

>  	case TC_HTB_DESTROY:
>  		return airoha_tc_htb_destroy(dev);
>  	case TC_HTB_NODE_MODIFY:
> diff --git a/drivers/net/ethernet/airoha/airoha_eth.h b/drivers/net/ethernet/airoha/airoha_eth.h
> index 8f42973f9cf5..8795af0010b6 100644
> --- a/drivers/net/ethernet/airoha/airoha_eth.h
> +++ b/drivers/net/ethernet/airoha/airoha_eth.h
> @@ -543,8 +543,8 @@ enum airoha_priv_flags {
>  };
>  
>  struct airoha_gdm_dev {
> +	struct airoha_qdma __rcu *qdma;
>  	struct airoha_gdm_port *port;
> -	struct airoha_qdma *qdma;
>  	struct airoha_eth *eth;
>  
>  	DECLARE_BITMAP(qos_sq_bmap, AIROHA_NUM_QOS_CHANNELS);
> @@ -676,7 +676,27 @@ int airoha_get_fe_port(struct airoha_gdm_dev *dev);
>  bool airoha_is_valid_gdm_dev(struct airoha_eth *eth,
>  			     struct airoha_gdm_dev *dev);
>  
> -void airoha_ppe_set_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport);
> +extern struct mutex flow_offload_mutex;
> +
> +static inline struct airoha_qdma *
> +airoha_qdma_deref(struct airoha_gdm_dev *dev)
> +{
> +	return rcu_dereference_protected(dev->qdma,
> +					 lockdep_rtnl_is_held() ||
> +					 lockdep_is_held(&flow_offload_mutex));
> +}
> +
> +void __airoha_ppe_set_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport);
> +void airoha_ppe_clear_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport);
> +
> +static inline void airoha_ppe_set_cpu_port(struct airoha_gdm_dev *dev,
> +					   u8 ppe_id, u8 fport)
> +{
> +	mutex_lock(&flow_offload_mutex);
> +	__airoha_ppe_set_cpu_port(dev, ppe_id, fport);
> +	mutex_unlock(&flow_offload_mutex);
> +}
> +
>  bool airoha_ppe_is_enabled(struct airoha_eth *eth, int index);
>  void airoha_ppe_check_skb(struct airoha_ppe_dev *dev, struct sk_buff *skb,
>  			  u16 hash, bool rx_wlan);
> diff --git a/drivers/net/ethernet/airoha/airoha_ppe.c b/drivers/net/ethernet/airoha/airoha_ppe.c
> index 91bcc55a6ac6..0ee0dd385645 100644
> --- a/drivers/net/ethernet/airoha/airoha_ppe.c
> +++ b/drivers/net/ethernet/airoha/airoha_ppe.c
> @@ -15,7 +15,7 @@
>  #include "airoha_regs.h"
>  #include "airoha_eth.h"
>  
> -static DEFINE_MUTEX(flow_offload_mutex);
> +DEFINE_MUTEX(flow_offload_mutex);
>  static DEFINE_SPINLOCK(ppe_lock);
>  
>  static const struct rhashtable_params airoha_flow_table_params = {
> @@ -84,10 +84,10 @@ static u32 airoha_ppe_get_timestamp(struct airoha_ppe *ppe)
>  			     AIROHA_FOE_IB1_BIND_TIMESTAMP);
>  }
>  
> -void airoha_ppe_set_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport)
> +void __airoha_ppe_set_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport)
>  {
> -	struct airoha_qdma *qdma = dev->qdma;
> -	struct airoha_eth *eth = qdma->eth;
> +	struct airoha_qdma *qdma = airoha_qdma_deref(dev);
> +	struct airoha_eth *eth = dev->eth;
>  	u8 qdma_id = qdma - &eth->qdma[0];
>  	u32 fe_cpu_port;
>  
> @@ -97,6 +97,14 @@ void airoha_ppe_set_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport)
>  		      __field_prep(DFT_CPORT_MASK(fport), fe_cpu_port));
>  }
>  
> +void airoha_ppe_clear_cpu_port(struct airoha_gdm_dev *dev, u8 ppe_id, u8 fport)
> +{
> +	mutex_lock(&flow_offload_mutex);
> +	airoha_fe_clear(dev->eth, REG_PPE_DFT_CPORT(ppe_id, fport),
> +			DFT_CPORT_MASK(fport));
> +	mutex_unlock(&flow_offload_mutex);
> +}
> +
>  static void airoha_ppe_hw_init(struct airoha_ppe *ppe)
>  {
>  	u32 sram_ppe_num_data_entries = PPE_SRAM_NUM_ENTRIES, sram_num_entries;
> @@ -195,7 +203,7 @@ static void airoha_ppe_hw_init(struct airoha_ppe *ppe)
>  			ppe_id = !airoha_is_lan_gdm_dev(dev) &&
>  				 airoha_ppe_is_enabled(eth, 1);
>  			fport = airoha_get_fe_port(dev);
> -			airoha_ppe_set_cpu_port(dev, ppe_id, fport);
> +			__airoha_ppe_set_cpu_port(dev, ppe_id, fport);
>  		}
>  	}
>  }
> diff --git a/drivers/net/ethernet/airoha/airoha_regs.h b/drivers/net/ethernet/airoha/airoha_regs.h
> index 436f3c8779c1..4e17dfbcf2b8 100644
> --- a/drivers/net/ethernet/airoha/airoha_regs.h
> +++ b/drivers/net/ethernet/airoha/airoha_regs.h
> @@ -376,6 +376,7 @@
>  
>  #define REG_SRC_PORT_FC_MAP6		0x2298
>  #define FC_ID_OF_SRC_PORT_MASK(_n)	GENMASK(4 + ((_n) << 3), ((_n) << 3))
> +#define FC_MAP6_DEF_VALUE		0x1b1a1918
>  
>  #define REG_CDM5_RX_OQ1_DROP_CNT	0x29d4
>  
> 
> -- 
> 2.54.0
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]