ethtool: Dynamic RSS context indirection table resizing

index 604966a398f5..84eb53b4172b 100644
@@ -8118,8 +8118,7 @@ static int __bnxt_reserve_rings(struct bnxt *bp)
 		    (bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) !=
 		     bnxt_get_nr_rss_ctxs(bp, rx_rings) ||
 		     bnxt_get_max_rss_ring(bp) >= rx_rings)) {
-			netdev_warn(bp->dev, "RSS table entries reverting to default\n");
-			bp->dev->priv_flags &= ~IFF_RXFH_CONFIGURED;
+			ethtool_rxfh_indir_lost(bp->dev);
 		}
 	}
 	bp->rx_nr_rings = rx_rings;

index f7009da94f0b..4429b4058daa 100644
@@ -6483,12 +6483,23 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv)
 
 	/* max number of channels may have changed */
 	max_nch = mlx5e_calc_max_nch(priv->mdev, priv->netdev, profile);
+
+	/* Locking is required by ethtool_rxfh_indir_lost() (sends
+	 * ETHTOOL_MSG_RSS_NTF) and by netif_set_real_num_*_queues in case
+	 * the netdev has been registered by this point (if this function
+	 * was called in the reload or resume flow).
+	 */
+	if (need_lock) {
+		rtnl_lock();
+		netdev_lock(priv->netdev);
+	}
+
 	if (priv->channels.params.num_channels > max_nch) {
 		mlx5_core_warn(priv->mdev, "MLX5E: Reducing number of channels to %d\n", max_nch);
 		/* Reducing the number of channels - RXFH has to be reset, and
 		 * mlx5e_num_channels_changed below will build the RQT.
 		 */
-		priv->netdev->priv_flags &= ~IFF_RXFH_CONFIGURED;
+		ethtool_rxfh_indir_lost(priv->netdev);
 		priv->channels.params.num_channels = max_nch;
 		if (priv->channels.params.mqprio.mode == TC_MQPRIO_MODE_CHANNEL) {
 			mlx5_core_warn(priv->mdev, "MLX5E: Disabling MQPRIO channel mode\n");

@@ -6505,15 +6516,7 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv)
 	/* 1. Set the real number of queues in the kernel the first time.
 	 * 2. Set our default XPS cpumask.
 	 * 3. Build the RQT.
-	 *
-	 * Locking is required by netif_set_real_num_*_queues in case the
-	 * netdev has been registered by this point (if this function was called
-	 * in the reload or resume flow).
 	 */
-	if (need_lock) {
-		rtnl_lock();
-		netdev_lock(priv->netdev);
-	}
 	err = mlx5e_num_channels_changed(priv);
 	if (need_lock) {
 		netdev_unlock(priv->netdev);

index 656d465bcd06..34ca9261de82 100644
@@ -176,6 +176,8 @@ static inline u32 ethtool_rxfh_indir_default(u32 index, u32 n_rx_rings)
  * struct ethtool_rxfh_context - a custom RSS context configuration
  * @indir_size: Number of u32 entries in indirection table
  * @key_size: Size of hash key, in bytes
+ * @indir_user_size: number of user provided entries for the
+ *	indirection table
  * @priv_size: Size of driver private data, in bytes
  * @hfunc: RSS hash function identifier.  One of the %ETH_RSS_HASH_*
  * @input_xfrm: Defines how the input data is transformed. Valid values are one

@@ -186,6 +188,7 @@ static inline u32 ethtool_rxfh_indir_default(u32 index, u32 n_rx_rings)
 struct ethtool_rxfh_context {
 	u32 indir_size;
 	u32 key_size;
+	u32 indir_user_size;
 	u16 priv_size;
 	u8 hfunc;
 	u8 input_xfrm;

@@ -214,6 +217,7 @@ static inline u8 *ethtool_rxfh_context_key(struct ethtool_rxfh_context *ctx)
 }
 
 void ethtool_rxfh_context_lost(struct net_device *dev, u32 context_id);
+void ethtool_rxfh_indir_lost(struct net_device *dev);
 
 struct link_mode_info {
 	int	speed;

@@ -1337,12 +1341,15 @@ int ethtool_virtdev_set_link_ksettings(struct net_device *dev,
  * @rss_ctx:		XArray of custom RSS contexts
  * @rss_lock:		Protects entries in @rss_ctx.  May be taken from
  *			within RTNL.
+ * @rss_indir_user_size: Number of user provided entries for the default
+ *			 (context 0) indirection table.
  * @wol_enabled:	Wake-on-LAN is enabled
  * @module_fw_flash_in_progress: Module firmware flashing is in progress.
  */
 struct ethtool_netdev_state {
 	struct xarray		rss_ctx;
 	struct mutex		rss_lock;
+	u32			rss_indir_user_size;
 	unsigned		wol_enabled:1;
 	unsigned		module_fw_flash_in_progress:1;
 };

index 7ca01eb3f7d2..e82de6831e05 100644
@@ -1716,7 +1716,6 @@ struct net_device_ops {
  * @IFF_OPENVSWITCH: device is a Open vSwitch master
  * @IFF_L3MDEV_SLAVE: device is enslaved to an L3 master device
  * @IFF_TEAM: device is a team device
- * @IFF_RXFH_CONFIGURED: device has had Rx Flow indirection table configured
  * @IFF_PHONY_HEADROOM: the headroom value is controlled by an external
  *	entity (i.e. the master device for bridged veth)
  * @IFF_MACSEC: device is a MACsec device

@@ -1752,7 +1751,6 @@ enum netdev_priv_flags {
 	IFF_OPENVSWITCH			= 1<<20,
 	IFF_L3MDEV_SLAVE		= 1<<21,
 	IFF_TEAM			= 1<<22,
-	IFF_RXFH_CONFIGURED		= 1<<23,
 	IFF_PHONY_HEADROOM		= 1<<24,
 	IFF_MACSEC			= 1<<25,
 	IFF_NO_RX_HANDLER		= 1<<26,

@@ -5569,10 +5567,7 @@ static inline bool netif_is_lag_port(const struct net_device *dev)
 	return netif_is_bond_slave(dev) || netif_is_team_port(dev);
 }
 
-static inline bool netif_is_rxfh_configured(const struct net_device *dev)
-{
-	return dev->priv_flags & IFF_RXFH_CONFIGURED;
-}
+bool netif_is_rxfh_configured(const struct net_device *dev);
 
 static inline bool netif_is_failover(const struct net_device *dev)
 {

index e252cf20c22f..d7d832fa9e00 100644
@@ -1204,6 +1204,34 @@ void ethtool_rxfh_context_lost(struct net_device *dev, u32 context_id)
 }
 EXPORT_SYMBOL(ethtool_rxfh_context_lost);
 
+bool netif_is_rxfh_configured(const struct net_device *dev)
+{
+	return dev->ethtool->rss_indir_user_size;
+}
+EXPORT_SYMBOL(netif_is_rxfh_configured);
+
+/**
+ * ethtool_rxfh_indir_lost - Notify core that the RSS indirection table was lost
+ * @dev: network device
+ *
+ * Drivers should call this when the device can no longer maintain the
+ * user-configured indirection table, typically after a HW fault recovery
+ * that reduced the maximum queue count. Marks the default RSS context
+ * indirection table as unconfigured and sends an %ETHTOOL_MSG_RSS_NTF
+ * notification.
+ */
+void ethtool_rxfh_indir_lost(struct net_device *dev)
+{
+	WARN_ONCE(!rtnl_is_locked() &&
+		  !lockdep_is_held_type(&dev->ethtool->rss_lock, -1),
+		  "RSS context lock assertion failed\n");
+
+	netdev_err(dev, "device error, RSS indirection table lost\n");
+	dev->ethtool->rss_indir_user_size = 0;
+	ethtool_rss_notify(dev, ETHTOOL_MSG_RSS_NTF, 0);
+}
+EXPORT_SYMBOL(ethtool_rxfh_indir_lost);
+
 enum ethtool_link_medium ethtool_str_to_medium(const char *str)
 {
 	int i;

index ff4b4780d6af..3d31a5a041e3 100644
@@ -1404,9 +1404,9 @@ static noinline_for_stack int ethtool_set_rxfh_indir(struct net_device *dev,
 
 	/* indicate whether rxfh was set to default */
 	if (user_size == 0)
-		dev->priv_flags &= ~IFF_RXFH_CONFIGURED;
+		dev->ethtool->rss_indir_user_size = 0;
 	else
-		dev->priv_flags |= IFF_RXFH_CONFIGURED;
+		dev->ethtool->rss_indir_user_size = rxfh_dev.indir_size;
 
 out_unlock:
 	mutex_unlock(&dev->ethtool->rss_lock);

@@ -1721,9 +1721,9 @@ static noinline_for_stack int ethtool_set_rxfh(struct net_device *dev,
 	if (!rxfh_dev.rss_context) {
 		/* indicate whether rxfh was set to default */
 		if (rxfh.indir_size == 0)
-			dev->priv_flags &= ~IFF_RXFH_CONFIGURED;
+			dev->ethtool->rss_indir_user_size = 0;
 		else if (rxfh.indir_size != ETH_RXFH_INDIR_NO_CHANGE)
-			dev->priv_flags |= IFF_RXFH_CONFIGURED;
+			dev->ethtool->rss_indir_user_size = dev_indir_size;
 	}
 	/* Update rss_ctx tracking */
 	if (rxfh_dev.rss_delete) {

@@ -1736,6 +1736,7 @@ static noinline_for_stack int ethtool_set_rxfh(struct net_device *dev,
 			ctx->indir_configured =
 				rxfh.indir_size &&
 				rxfh.indir_size != ETH_RXFH_INDIR_NO_CHANGE;
+			ctx->indir_user_size = dev_indir_size;
 		}
 		if (rxfh_dev.key) {
 			memcpy(ethtool_rxfh_context_key(ctx), rxfh_dev.key,

index da5934cceb07..e6fc6e64fb27 100644
@@ -686,7 +686,7 @@ rss_set_prep_indir(struct net_device *dev, struct genl_info *info,
 
 	*mod |= memcmp(rxfh->indir, data->indir_table, data->indir_size);
 
-	return 0;
+	return user_size;
 
 err_free:
 	kfree(rxfh->indir);

@@ -833,6 +833,7 @@ ethnl_rss_set(struct ethnl_req_info *req_info, struct genl_info *info)
 	struct nlattr **tb = info->attrs;
 	struct rss_reply_data data = {};
 	const struct ethtool_ops *ops;
+	u32 indir_user_size;
 	int ret;
 
 	ops = dev->ethtool_ops;

@@ -845,8 +846,9 @@ ethnl_rss_set(struct ethnl_req_info *req_info, struct genl_info *info)
 	rxfh.rss_context = request->rss_context;
 
 	ret = rss_set_prep_indir(dev, info, &data, &rxfh, &indir_reset, &mod);
-	if (ret)
+	if (ret < 0)
 		goto exit_clean_data;
+	indir_user_size = ret;
 	indir_mod = !!tb[ETHTOOL_A_RSS_INDIR];
 
 	rxfh.hfunc = data.hfunc;

@@ -889,12 +891,15 @@ ethnl_rss_set(struct ethnl_req_info *req_info, struct genl_info *info)
 	if (ret)
 		goto exit_unlock;
 
-	if (ctx)
+	if (ctx) {
 		rss_set_ctx_update(ctx, tb, &data, &rxfh);
-	else if (indir_reset)
-		dev->priv_flags &= ~IFF_RXFH_CONFIGURED;
-	else if (indir_mod)
-		dev->priv_flags |= IFF_RXFH_CONFIGURED;
+		if (indir_user_size)
+			ctx->indir_user_size = indir_user_size;
+	} else if (indir_reset) {
+		dev->ethtool->rss_indir_user_size = 0;
+	} else if (indir_mod) {
+		dev->ethtool->rss_indir_user_size = indir_user_size;
+	}
 
 exit_unlock:
 	mutex_unlock(&dev->ethtool->rss_lock);

@@ -999,6 +1004,7 @@ int ethnl_rss_create_doit(struct sk_buff *skb, struct genl_info *info)
 	const struct ethtool_ops *ops;
 	struct rss_req_info req = {};
 	struct net_device *dev;
+	u32 indir_user_size;
 	struct sk_buff *rsp;
 	void *hdr;
 	u32 limit;

@@ -1035,8 +1041,9 @@ int ethnl_rss_create_doit(struct sk_buff *skb, struct genl_info *info)
 		goto exit_ops;
 
 	ret = rss_set_prep_indir(dev, info, &data, &rxfh, &indir_dflt, &mod);
-	if (ret)
+	if (ret < 0)
 		goto exit_clean_data;
+	indir_user_size = ret;
 
 	ethnl_update_u8(&rxfh.hfunc, tb[ETHTOOL_A_RSS_HFUNC], &mod);
 

@@ -1080,6 +1087,7 @@ int ethnl_rss_create_doit(struct sk_buff *skb, struct genl_info *info)
 
 	/* Store the config from rxfh to Xarray.. */
 	rss_set_ctx_update(ctx, tb, &data, &rxfh);
+	ctx->indir_user_size = indir_user_size;
 	/* .. copy from Xarray to data. */
 	__rss_prepare_ctx(dev, &data, ctx);
 

index 34ca9261de82..1cb0740ba331 100644
@@ -218,6 +218,12 @@ static inline u8 *ethtool_rxfh_context_key(struct ethtool_rxfh_context *ctx)
 
 void ethtool_rxfh_context_lost(struct net_device *dev, u32 context_id);
 void ethtool_rxfh_indir_lost(struct net_device *dev);
+bool ethtool_rxfh_indir_can_resize(struct net_device *dev, const u32 *tbl,
+				   u32 old_size, u32 new_size);
+void ethtool_rxfh_indir_resize(struct net_device *dev, u32 *tbl,
+			       u32 old_size, u32 new_size);
+int ethtool_rxfh_ctxs_can_resize(struct net_device *dev, u32 new_indir_size);
+void ethtool_rxfh_ctxs_resize(struct net_device *dev, u32 new_indir_size);
 
 struct link_mode_info {
 	int	speed;

161 lines not on this page · raw message

index 48e8e3be70d3..b87ac2bb43dd 100644
@@ -942,6 +942,7 @@ static int bnxt_set_channels(struct net_device *dev,
 {
 	struct bnxt *bp = netdev_priv(dev);
 	int req_tx_rings, req_rx_rings, tcs;
+	u32 new_tbl_size = 0, old_tbl_size;
 	bool sh = false;
 	int tx_xdp = 0;
 	int rc = 0;

@@ -977,19 +978,33 @@ static int bnxt_set_channels(struct net_device *dev,
 		tx_xdp = req_rx_rings;
 	}
 
-	if (bnxt_get_nr_rss_ctxs(bp, req_rx_rings) !=
-	    bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings) &&
-	    (netif_is_rxfh_configured(dev) || bp->num_rss_ctx)) {
-		netdev_warn(dev, "RSS table size change required, RSS table entries must be default (with no additional RSS contexts present) to proceed\n");
-		return -EINVAL;
-	}
-
 	rc = bnxt_check_rings(bp, req_tx_rings, req_rx_rings, sh, tcs, tx_xdp);
 	if (rc) {
 		netdev_warn(dev, "Unable to allocate the requested rings\n");
 		return rc;
 	}
 
+	/* RSS table size only changes on P5 chips with older firmware;
+	 * newer firmware always uses the largest table size.
+	 */
+	if (bnxt_get_nr_rss_ctxs(bp, req_rx_rings) !=
+	    bnxt_get_nr_rss_ctxs(bp, bp->rx_nr_rings)) {
+		new_tbl_size = bnxt_get_nr_rss_ctxs(bp, req_rx_rings) *
+			       BNXT_RSS_TABLE_ENTRIES_P5;
+		old_tbl_size = bnxt_get_rxfh_indir_size(dev);
+
+		if (!ethtool_rxfh_indir_can_resize(dev, bp->rss_indir_tbl,
+						   old_tbl_size,
+						   new_tbl_size)) {
+			netdev_warn(dev, "RSS table resize not possible\n");
+			return -EINVAL;
+		}
+
+		rc = ethtool_rxfh_ctxs_can_resize(dev, new_tbl_size);
+		if (rc)
+			return rc;
+	}
+
 	if (netif_running(dev)) {
 		if (BNXT_PF(bp)) {
 			/* TODO CHIMP_FW: Send message to all VF's

@@ -999,6 +1014,12 @@ static int bnxt_set_channels(struct net_device *dev,
 		bnxt_close_nic(bp, true, false);
 	}
 
+	if (new_tbl_size) {
+		ethtool_rxfh_indir_resize(dev, bp->rss_indir_tbl,
+					  old_tbl_size, new_tbl_size);
+		ethtool_rxfh_ctxs_resize(dev, new_tbl_size);
+	}
+
 	if (sh) {
 		bp->flags |= BNXT_FLAG_SHARED_RINGS;
 		bp->rx_nr_rings = channel->combined_count;