net: ethernet: mtk_eth_soc: account for vlan in rx header length
Subsystems (6)
Felix Fietkau
maintainerPATCH v3 1/5 net: ethernet: mtk_eth_soc: account for vlan in rx header length
The network stack assumes that devices can handle an extra VLAN tag without increasing the MTU
--- drivers/net/ethernet/mediatek/mtk_eth_soc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.h +1 -1
@@ -34,7 +34,7 @@#define MTK_QDMA_RING_SIZE 2048#define MTK_DMA_SIZE 512#define MTK_MAC_COUNT 2 -#define MTK_RX_ETH_HLEN (ETH_HLEN + ETH_FCS_LEN) +#define MTK_RX_ETH_HLEN (VLAN_ETH_HLEN + ETH_FCS_LEN)#define MTK_RX_HLEN (NET_SKB_PAD + MTK_RX_ETH_HLEN + NET_IP_ALIGN)#define MTK_DMA_DUMMY_DESC 0xffffffff#define MTK_DEFAULT_MSG_ENABLE (NETIF_MSG_DRV | \
Signature · 1 line
2.38.1
Felix Fietkau
maintainerPATCH v3 2/5 net: ethernet: mtk_eth_soc: work around issue with sending small fragments
When lots of frames are sent with a number of very small fragments, an internal FIFO can overflow, causing the DMA engine to lock up lock up and transmit attempts time out.
Fix this on MT7986 by increasing the reserved FIFO space.
Fix this on older chips by detecting the presence of small fragments and use skb_gso_segment + skb_linearize to deal with them.
---
v2: add proper fix for MT7986, limit workaround to NETSYS v1
v3: fix a regression on mt7986
drivers/net/ethernet/mediatek/mtk_eth_soc.c | 38 +++++++++++++++++++-- drivers/net/ethernet/mediatek/mtk_eth_soc.h | 2 +- 2 files changed, 37 insertions(+), 3 deletions(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +36 -2
@@ -1441,12 +1441,28 @@ static void mtk_wake_queue(struct mtk_eth *eth) } } +static bool mtk_skb_has_small_frag (struct sk_buff *skb) +{ +int min_size =16 ; +int i; + +if (skb_headlen (skb) < min_size) +return true; + +for (i =0 ; i <skb_shinfo (skb)->nr_frags; i++) +if (skb_frag_size (&skb_shinfo (skb)->frags[i]) < min_size) +return true; + +return false; +} +static netdev_tx_t mtk_start_xmit (struct sk_buff *skb,struct net_device *dev) {struct mtk_mac *mac =netdev_priv (dev);struct mtk_eth *eth = mac->hw;struct mtk_tx_ring *ring = ð->tx_ring;struct net_device_stats *stats = &dev->stats; +struct sk_buff *segs, *next;bool gso = false;int tx_num;
@@ -1468,6 +1484,18 @@ static netdev_tx_t mtk_start_xmit(struct sk_buff *skb, struct net_device *dev)return NETDEV_TX_BUSY ; } +if (!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 ) && +skb_is_gso (skb) &&mtk_skb_has_small_frag (skb)) { + segs =skb_gso_segment (skb, dev->features & ~NETIF_F_ALL_TSO ); +if (IS_ERR (segs)) + goto drop; + +if (segs) { +consume_skb (skb); + skb = segs; + } + } +/* TSO: fill MSS info in tcp checksum field */ if (skb_is_gso (skb)) {if (skb_cow_head (skb,0 )) {
@@ -1483,8 +1511,14 @@ static netdev_tx_t mtk_start_xmit(struct sk_buff *skb, struct net_device *dev) } } -if (mtk_tx_map (skb, dev, tx_num, ring, gso) <0 ) - goto drop; + skb_list_walk_safe(skb, skb, next) { +if ((!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 ) && +mtk_skb_has_small_frag (skb) &&skb_linearize (skb)) || +mtk_tx_map (skb, dev, tx_num, ring, gso) <0 ) { + stats->tx_dropped++; +dev_kfree_skb_any (skb); + } + }if (unlikely (atomic_read (&ring->free_count) <= ring->thresh))netif_tx_stop_all_queues (dev);
drivers/net/ethernet/mediatek/mtk_eth_soc.h +1 -1
@@ -247,7 +247,7 @@#define MTK_CHK_DDONE_EN BIT(28)#define MTK_DMAD_WR_WDONE BIT(26)#define MTK_WCOMP_EN BIT(24) -#define MTK_RESV_BUF (0x40 << 16) +#define MTK_RESV_BUF (0x80 << 16)#define MTK_MUTLI_CNT (0x4 << 12)#define MTK_LEAKY_BUCKET_EN BIT(11)
Signature · 1 line
2.38.1
Felix Fietkau
maintainerPATCH v3 3/5 net: ethernet: mtk_eth_soc: fix flow_offload related refcount bug
In reply to Felix FietkauSince we call flow_block_cb_decref on FLOW_BLOCK_UNBIND, we need to call flow_block_cb_incref unconditionally, even for a newly allocated cb. Fixes a use-after-free bug
--- .../net/ethernet/mediatek/mtk_ppe_offload.c | 21 ++++++++++++------- 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/net/ethernet/mediatek/mtk_ppe_offload.c +13 -8
@@ -554,6 +554,7 @@ mtk_eth_setup_tc_block(struct net_device *dev, struct flow_block_offload *f)struct mtk_eth *eth = mac->hw;static LIST_HEAD (block_cb_list);struct flow_block_cb *block_cb; +bool register_block = false;flow_setup_cb_t *cb;if (!eth->soc->offload_version)
@@ -568,16 +569,20 @@ mtk_eth_setup_tc_block(struct net_device *dev, struct flow_block_offload *f)switch (f->command) {case FLOW_BLOCK_BIND : block_cb =flow_block_cb_lookup (f->block, cb, dev); -if (block_cb) { -flow_block_cb_incref (block_cb); -return 0 ; +if (!block_cb) { + block_cb =flow_block_cb_alloc (cb, dev, dev,NULL ); +if (IS_ERR (block_cb)) +return PTR_ERR (block_cb); + + register_block = true; } - block_cb =flow_block_cb_alloc (cb, dev, dev,NULL ); -if (IS_ERR (block_cb)) -return PTR_ERR (block_cb); -flow_block_cb_add (block_cb, f); -list_add_tail (&block_cb->driver_list, &block_cb_list); +flow_block_cb_incref (block_cb); + +if (register_block) { +flow_block_cb_add (block_cb, f); +list_add_tail (&block_cb->driver_list, &block_cb_list); + }return 0 ;case FLOW_BLOCK_UNBIND : block_cb =flow_block_cb_lookup (f->block, cb, dev);
Signature · 1 line
2.38.1
Felix Fietkau
maintainerPATCH v3 4/5 net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
In reply to Felix FietkauThrough testing I found out that hardware vlan rx offload support seems to have some hardware issues. At least when using multiple MACs and when receiving tagged packets on the secondary MAC, the hardware can sometimes start to emit wrong tags on the first MAC as well.
In order to avoid such issues, drop the feature configuration and use the offload feature only for DSA hardware untagging on MT7621/MT7622 devices which only use one MAC.
---
v3: rebase
drivers/net/ethernet/mediatek/mtk_eth_soc.c | 101 ++++++++------------ drivers/net/ethernet/mediatek/mtk_eth_soc.h | 1 - 2 files changed, 39 insertions(+), 63 deletions(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +39 -62
@@ -2042,29 +2042,16 @@ static int mtk_poll_rx(struct napi_struct *napi, int budget,if (reason ==MTK_PPE_CPU_REASON_HIT_UNBIND_RATE_REACHED )mtk_ppe_check_skb (eth->ppe[0 ], skb, hash); -if (netdev->features &NETIF_F_HW_VLAN_CTAG_RX ) { -if (MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 )) { -if (trxd.rxd3 &RX_DMA_VTAG_V2 ) -__vlan_hwaccel_put_tag (skb, -htons (RX_DMA_VPID (trxd.rxd4)), -RX_DMA_VID (trxd.rxd4)); - }else if (trxd.rxd2 &RX_DMA_VTAG ) { -__vlan_hwaccel_put_tag (skb,htons (RX_DMA_VPID (trxd.rxd3)), -RX_DMA_VID (trxd.rxd3)); - } - } -/* When using VLAN untagging in combination with DSA, the * hardware treats the MTK special tag as a VLAN and untags it. */ -if (skb_vlan_tag_present (skb) &&netdev_uses_dsa (netdev)) { -unsigned int port =ntohs (skb->vlan_proto) &GENMASK (2 ,0 ); +if (!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 ) && + (trxd.rxd2 &RX_DMA_VTAG ) &&netdev_uses_dsa (netdev)) { +unsigned int port =RX_DMA_VPID (trxd.rxd3) &GENMASK (2 ,0 );if (port <ARRAY_SIZE (eth->dsa_meta) && eth->dsa_meta[port])skb_dst_set_noref (skb, ð->dsa_meta[port]->dst); - -__vlan_hwaccel_clear_tag (skb); }skb_record_rx_queue (skb,0 );
@@ -2889,29 +2876,11 @@ static netdev_features_t mtk_fix_features(struct net_device *dev,static int mtk_set_features (struct net_device *dev,netdev_features_t features) { -struct mtk_mac *mac =netdev_priv (dev); -struct mtk_eth *eth = mac->hw;netdev_features_t diff = dev->features ^ features; -int i;if ((diff &NETIF_F_LRO ) && !(features &NETIF_F_LRO ))mtk_hwlro_netdev_disable (dev); -/* Set RX VLAN offloading */ -if (!(diff &NETIF_F_HW_VLAN_CTAG_RX )) -return 0 ; - -mtk_w32 (eth, !!(features &NETIF_F_HW_VLAN_CTAG_RX ), -MTK_CDMP_EG_CTRL ); - -/* sync features with other MAC */ -for (i =0 ; i <MTK_MAC_COUNT ; i++) { -if (!eth->netdev[i] || eth->netdev[i] == dev) -continue ; - eth->netdev[i]->features &= ~NETIF_F_HW_VLAN_CTAG_RX ; - eth->netdev[i]->features |= features &NETIF_F_HW_VLAN_CTAG_RX ; - } -return 0 ; }
@@ -3211,30 +3180,6 @@ static int mtk_open(struct net_device *dev)struct mtk_eth *eth = mac->hw;int i, err; -if (mtk_uses_dsa (dev) && !eth->prog) { -for (i =0 ; i <ARRAY_SIZE (eth->dsa_meta); i++) { -struct metadata_dst *md_dst = eth->dsa_meta[i]; - -if (md_dst) -continue ; - - md_dst =metadata_dst_alloc (0 ,METADATA_HW_PORT_MUX , -GFP_KERNEL ); -if (!md_dst) -return -ENOMEM ; - - md_dst->u.port_info.port_id = i; - eth->dsa_meta[i] = md_dst; - } - }else { -/* Hardware special tag parsing needs to be disabled if at least -* one MAC does not use DSA. -*/ -u32 val =mtk_r32 (eth,MTK_CDMP_IG_CTRL ); - val &= ~MTK_CDMP_STAG_EN ; -mtk_w32 (eth, val,MTK_CDMP_IG_CTRL ); - } - err =phylink_of_phy_connect (mac->phylink, mac->of_node,0 );if (err) { netdev_err(dev, "%s: could not attach PHY: %d\n", __func__,
@@ -3273,6 +3218,39 @@ static int mtk_open(struct net_device *dev)phylink_start (mac->phylink);netif_tx_start_all_queues (dev); +if (MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 )) +return 0 ; + +if (mtk_uses_dsa (dev) && !eth->prog) { +for (i =0 ; i <ARRAY_SIZE (eth->dsa_meta); i++) { +struct metadata_dst *md_dst = eth->dsa_meta[i]; + +if (md_dst) +continue ; + + md_dst =metadata_dst_alloc (0 ,METADATA_HW_PORT_MUX , +GFP_KERNEL ); +if (!md_dst) +return -ENOMEM ; + + md_dst->u.port_info.port_id = i; + eth->dsa_meta[i] = md_dst; + } + }else { +/* Hardware special tag parsing needs to be disabled if at least +* one MAC does not use DSA. +*/ +u32 val =mtk_r32 (eth,MTK_CDMP_IG_CTRL ); + val &= ~MTK_CDMP_STAG_EN ; +mtk_w32 (eth, val,MTK_CDMP_IG_CTRL ); + + val =mtk_r32 (eth,MTK_CDMQ_IG_CTRL ); + val &= ~MTK_CDMQ_STAG_EN ; +mtk_w32 (eth, val,MTK_CDMQ_IG_CTRL ); + +mtk_w32 (eth,0 ,MTK_CDMP_EG_CTRL ); + } +return 0 ; }
@@ -3599,10 +3577,9 @@ static int mtk_hw_init(struct mtk_eth *eth)if (!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 )) { val =mtk_r32 (eth,MTK_CDMP_IG_CTRL );mtk_w32 (eth, val |MTK_CDMP_STAG_EN ,MTK_CDMP_IG_CTRL ); - } -/* Enable RX VLan Offloading */ -mtk_w32 (eth,1 ,MTK_CDMP_EG_CTRL ); +mtk_w32 (eth,1 ,MTK_CDMP_EG_CTRL ); + }/* set interrupt delays based on current Net DIM sample */ mtk_dim_rx (ð->rx_dim.work);
@@ -4203,7 +4180,7 @@ static int mtk_add_mac(struct mtk_eth *eth, struct device_node *np) eth->netdev[id]->hw_features |=NETIF_F_LRO ; eth->netdev[id]->vlan_features = eth->soc->hw_features & - ~(NETIF_F_HW_VLAN_CTAG_TX |NETIF_F_HW_VLAN_CTAG_RX ); + ~NETIF_F_HW_VLAN_CTAG_TX ; eth->netdev[id]->features |= eth->soc->hw_features; eth->netdev[id]->ethtool_ops = &mtk_ethtool_ops;
drivers/net/ethernet/mediatek/mtk_eth_soc.h +0 -1
@@ -48,7 +48,6 @@#define MTK_HW_FEATURES (NETIF_F_IP_CSUM | \ NETIF_F_RXCSUM | \ NETIF_F_HW_VLAN_CTAG_TX | \ - NETIF_F_HW_VLAN_CTAG_RX | \ NETIF_F_SG | NETIF_F_TSO | \ NETIF_F_TSO6 | \ NETIF_F_IPV6_CSUM |\
Signature · 1 line
2.38.1
Aw: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Hi,
v2 or v3 seems to break vlan on mt7986 over eth0 (mt7531 switch). v1 was working on next from end of November. But my rebased tree with v1 on 6.2-rc1 has same issue, so something after next 2711 was added which break vlan over mt7531.
Directly over eth1 it works (was not working before).
if i made no mistake there is still something wrong.
btw. mt7622/r64 can also use second gmac (over vlan aware bridge with aux-port of switch to wan-port) it is only not default in mainline. But maybe this should not be used as decision for dropping "dsa-tag" (wrongly vlan-tag).
regards Frank
Felix Fietkau
maintainerRe: Aw: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Quoted 11 lines · Frank Wunderlich
Hi,
v2 or v3 seems to break vlan on mt7986 over eth0 (mt7531 switch). v1 was working on next from end of November. But my rebased tree with v1 on 6.2-rc1 has same issue, so something after next 2711 was added which break vlan over mt7531.
Directly over eth1 it works (was not working before).
if i made no mistake there is still something wrong.
btw. mt7622/r64 can also use second gmac (over vlan aware bridge with aux-port of switch to wan-port) it is only not default in mainline. But maybe this should not be used as decision for dropping "dsa-tag" (wrongly vlan-tag).
regards Frank
Thanks for reporting.
Please try this patch on top of the series:
---
drivers/net/ethernet/mediatek/mtk_eth_soc.c +2 -8
@@ -3218,10 +3218,8 @@ static int mtk_open(struct net_device *dev)phylink_start (mac->phylink);netif_tx_start_all_queues (dev); -if (MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 )) -return 0 ; - -if (mtk_uses_dsa (dev) && !eth->prog) { +if (!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 ) && +mtk_uses_dsa (dev) && !eth->prog) {for (i =0 ; i <ARRAY_SIZE (eth->dsa_meta); i++) {struct metadata_dst *md_dst = eth->dsa_meta[i];
@@ -3244,10 +3242,6 @@ static int mtk_open(struct net_device *dev) val &= ~MTK_CDMP_STAG_EN ;mtk_w32 (eth, val,MTK_CDMP_IG_CTRL ); - val =mtk_r32 (eth,MTK_CDMQ_IG_CTRL ); - val &= ~MTK_CDMQ_STAG_EN ; -mtk_w32 (eth, val,MTK_CDMQ_IG_CTRL ); -mtk_w32 (eth,0 ,MTK_CDMP_EG_CTRL ); }
Aw: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Hi
thanks for fast response
Quoted 39 lines
Gesendet: Freitag, 30. Dezember 2022 um 13:56 Uhr
Von: "Felix Fietkau" <[email protected]>
An: "Frank Wunderlich" <[email protected]>
Cc: [email protected], "John Crispin" <[email protected]>, "Sean Wang" <[email protected]>, "Mark Lee" <[email protected]>, "Lorenzo Bianconi" <[email protected]>, "David S. Miller" <[email protected]>, "Eric Dumazet" <[email protected]>, "Jakub Kicinski" <[email protected]>, "Paolo Abeni" <[email protected]>, "Matthias Brugger" <[email protected]>, "Russell King" <[email protected]>, [email protected], [email protected], [email protected]
Betreff: Re: Aw: [PATCH net v3 4/5] net: ethernet:mtk_eth_soc: drop generic vlan rx offload, only use DSA untaggingQuoted 11 lines · Frank Wunderlich
Hi,
v2 or v3 seems to break vlan on mt7986 over eth0 (mt7531 switch). v1 was working on next from end of November. But my rebased tree with v1 on 6.2-rc1 has same issue, so something after next 2711 was added which break vlan over mt7531.
Directly over eth1 it works (was not working before).
if i made no mistake there is still something wrong.
btw. mt7622/r64 can also use second gmac (over vlan aware bridge with aux-port of switch to wan-port) it is only not default in mainline. But maybe this should not be used as decision for dropping "dsa-tag" (wrongly vlan-tag).
regards Frank
Thanks for reporting.
Please try this patch on top of the series:
---drivers/net/ethernet/mediatek/mtk_eth_soc.c +2
-4@@ -3218,10 +3218,8 @@ static int mtk_open(struct net_device *dev)phylink_start (mac->phylink);netif_tx_start_all_queues (dev); -if (MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 )) -return 0 ; - -if (mtk_uses_dsa (dev) && !eth->prog) { +if (!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 ) && +mtk_uses_dsa (dev) && !eth->prog) {for (i =0 ; i <ARRAY_SIZE (eth->dsa_meta); i++) {struct metadata_dst *md_dst = eth->dsa_meta[i];
drivers/net/ethernet/mediatek/mtk_eth_soc.c static int mtk_open(struct net_device *dev) · line 3247
Quoted 10 lines
drivers/net/ethernet/mediatek/mtk_eth_soc.c +0
-4@@ -3244,10 +3242,6 @@ static int mtk_open(struct net_device *dev) val &= ~MTK_CDMP_STAG_EN ;mtk_w32 (eth, val,MTK_CDMP_IG_CTRL ); - val =mtk_r32 (eth,MTK_CDMQ_IG_CTRL ); - val &= ~MTK_CDMQ_STAG_EN ; -mtk_w32 (eth, val,MTK_CDMQ_IG_CTRL ); -mtk_w32 (eth,0 ,MTK_CDMP_EG_CTRL ); }
seems not helping...this is how i test it:
ip link set eth1 up
ip link add link eth1 name vlan500 type vlan id 500 ip link add link wan name vlan600 type vlan id 600 ip addr add 192.168.50.1/24 dev vlan500
ip addr add 192.168.60.1/24 dev vlan600
ip link set vlan500 up
ip link set wan up
ip link set vlan600 up
#do this on the other side:
#netif=enp3s0
#sudo ip link add link $netif name vlan500 type vlan id 500 #sudo ip link add link $netif name vlan600 type vlan id 600 #sudo ip link set vlan500 up
#sudo ip link set vlan600 up
#sudo ip addr add 192.168.50.2/24 dev vlan500 #sudo ip addr add 192.168.60.2/24 dev vlan600
verified all used ports on my switch are in trunk-mode with vlan-membership of these 2 vlan.
booted 6.1 and there is vlan on dsa-port broken too, so either my test-setup is broken or code...but wonder why 6.1 is broken too...
with tcp-dump on my laptop i see that some packets came in for both vlan, but they seem not valid, arp only for vlan 500 (eth1 on r3).
$ sudo tcpdump -i enp3s0 -nn -e vlan tcpdump: verbose output suppressed, use -v[v]... for full protocol decode listening on enp3s0, link-type EN10MB (Ethernet), snapshot length 262144 bytes 14:49:09.548909 e4:b9:7a:f7:c4:8b > 33:33:00:00:83:84, ethertype 802.1Q (0x8100), length 524: vlan 500, p 0, ethertype IPv6 (0x86dd), fe80::e6b9:7aff:fef7:c48b.34177 > ff12::8384.21027: UDP, length 458 14:49:09.548929 e4:b9:7a:f7:c4:8b > 33:33:00:00:83:84, ethertype 802.1Q (0x8100), length 524: vlan 600, p 0, ethertype IPv6 (0x86dd), fe80::e6b9:7aff:fef7:c48b.34177 > ff12::8384.21027: UDP, length 458 14:49:09.549470 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 504: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2.33050 > 192.168.50.255.21027: UDP, length 458 14:49:09.549522 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 504: vlan 600, p 0, ethertype IPv4 (0x0800), 192.168.60.2.33050 > 192.168.60.255.21027: UDP, length 458 14:49:26.324503 92:65:f3:ec:b0:19 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 60: vlan 500, p 0, ethertype ARP (0x0806), Request who-has 192.168.50.2 tell 192.168.50.1, length 42 14:49:26.324525 e4:b9:7a:f7:c4:8b > 92:65:f3:ec:b0:19, ethertype 802.1Q (0x8100), length 46: vlan 500, p 0, ethertype ARP (0x0806), Reply 192.168.50.2 is-at e4:b9:7a:f7:c4:8b, length 28 14:49:26.325091 92:65:f3:ec:b0:19 > e4:b9:7a:f7:c4:8b, ethertype 802.1Q (0x8100), length 102: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.1 > 192.168.50.2: ICMP echo request, id 44246, seq 1, length 64 14:49:26.325158 e4:b9:7a:f7:c4:8b > 92:65:f3:ec:b0:19, ethertype 802.1Q (0x8100), length 102: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2 > 192.168.50.1: ICMP echo reply, id 44246, seq 1, length 64
on r3 i see these packets going out (so far it looks good):
root@bpi-r3:~# ping 192.168.60.2 PING 192.168.60.2 (192.168.60.2) 56(84) bytes of data. 13:30:29.782003 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 13:30:30.788175 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 13:30:31.828181 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 From 192.168.60.1 icmp_seq=1 Destination Host Unreachable 13:30:32.868205 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 13:30:33.908171 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28
HTH, maybe daniel or anyone other can confirm this
regards Frank
Felix Fietkau
maintainerRe: Aw: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Quoted 110 lines · Frank Wunderlich
Hi
thanks for fast response
Quoted 39 lines
Gesendet: Freitag, 30. Dezember 2022 um 13:56 Uhr
Von: "Felix Fietkau" <[email protected]>
An: "Frank Wunderlich" <[email protected]>
Cc: [email protected], "John Crispin" <[email protected]>, "Sean Wang" <[email protected]>, "Mark Lee" <[email protected]>, "Lorenzo Bianconi" <[email protected]>, "David S. Miller" <[email protected]>, "Eric Dumazet" <[email protected]>, "Jakub Kicinski" <[email protected]>, "Paolo Abeni" <[email protected]>, "Matthias Brugger" <[email protected]>, "Russell King" <[email protected]>, [email protected], [email protected], [email protected]
Betreff: Re: Aw: [PATCH net v3 4/5] net: ethernet:mtk_eth_soc: drop generic vlan rx offload, only use DSA untaggingQuoted 11 lines · Frank Wunderlich
Hi,
v2 or v3 seems to break vlan on mt7986 over eth0 (mt7531 switch). v1 was working on next from end of November. But my rebased tree with v1 on 6.2-rc1 has same issue, so something after next 2711 was added which break vlan over mt7531.
Directly over eth1 it works (was not working before).
if i made no mistake there is still something wrong.
btw. mt7622/r64 can also use second gmac (over vlan aware bridge with aux-port of switch to wan-port) it is only not default in mainline. But maybe this should not be used as decision for dropping "dsa-tag" (wrongly vlan-tag).
regards Frank
Thanks for reporting.
Please try this patch on top of the series:
---drivers/net/ethernet/mediatek/mtk_eth_soc.c +2
-4@@ -3218,10 +3218,8 @@ static int mtk_open(struct net_device *dev)phylink_start (mac->phylink);netif_tx_start_all_queues (dev); -if (MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 )) -return 0 ; - -if (mtk_uses_dsa (dev) && !eth->prog) { +if (!MTK_HAS_CAPS (eth->soc->caps,MTK_NETSYS_V2 ) && +mtk_uses_dsa (dev) && !eth->prog) {for (i =0 ; i <ARRAY_SIZE (eth->dsa_meta); i++) {struct metadata_dst *md_dst = eth->dsa_meta[i];drivers/net/ethernet/mediatek/mtk_eth_soc.c static int mtk_open(struct net_device *dev) · line 3247
Quoted 10 lines
drivers/net/ethernet/mediatek/mtk_eth_soc.c +0
-4@@ -3244,10 +3242,6 @@ static int mtk_open(struct net_device *dev) val &= ~MTK_CDMP_STAG_EN ;mtk_w32 (eth, val,MTK_CDMP_IG_CTRL ); - val =mtk_r32 (eth,MTK_CDMQ_IG_CTRL ); - val &= ~MTK_CDMQ_STAG_EN ; -mtk_w32 (eth, val,MTK_CDMQ_IG_CTRL ); -mtk_w32 (eth,0 ,MTK_CDMP_EG_CTRL ); }seems not helping...this is how i test it:
ip link set eth1 up
ip link add link eth1 name vlan500 type vlan id 500 ip link add link wan name vlan600 type vlan id 600 ip addr add 192.168.50.1/24 dev vlan500
ip addr add 192.168.60.1/24 dev vlan600
ip link set vlan500 up
ip link set wan up
ip link set vlan600 up#do this on the other side:
#netif=enp3s0
#sudo ip link add link $netif name vlan500 type vlan id 500 #sudo ip link add link $netif name vlan600 type vlan id 600 #sudo ip link set vlan500 up
#sudo ip link set vlan600 up
#sudo ip addr add 192.168.50.2/24 dev vlan500 #sudo ip addr add 192.168.60.2/24 dev vlan600verified all used ports on my switch are in trunk-mode with vlan-membership of these 2 vlan.
booted 6.1 and there is vlan on dsa-port broken too, so either my test-setup is broken or code...but wonder why 6.1 is broken too...
with tcp-dump on my laptop i see that some packets came in for both vlan, but they seem not valid, arp only for vlan 500 (eth1 on r3).
$ sudo tcpdump -i enp3s0 -nn -e vlan tcpdump: verbose output suppressed, use -v[v]... for full protocol decode listening on enp3s0, link-type EN10MB (Ethernet), snapshot length 262144 bytes 14:49:09.548909 e4:b9:7a:f7:c4:8b > 33:33:00:00:83:84, ethertype 802.1Q (0x8100), length 524: vlan 500, p 0, ethertype IPv6 (0x86dd), fe80::e6b9:7aff:fef7:c48b.34177 > ff12::8384.21027: UDP, length 458 14:49:09.548929 e4:b9:7a:f7:c4:8b > 33:33:00:00:83:84, ethertype 802.1Q (0x8100), length 524: vlan 600, p 0, ethertype IPv6 (0x86dd), fe80::e6b9:7aff:fef7:c48b.34177 > ff12::8384.21027: UDP, length 458 14:49:09.549470 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 504: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2.33050 > 192.168.50.255.21027: UDP, length 458 14:49:09.549522 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 504: vlan 600, p 0, ethertype IPv4 (0x0800), 192.168.60.2.33050 > 192.168.60.255.21027: UDP, length 458 14:49:26.324503 92:65:f3:ec:b0:19 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 60: vlan 500, p 0, ethertype ARP (0x0806), Request who-has 192.168.50.2 tell 192.168.50.1, length 42 14:49:26.324525 e4:b9:7a:f7:c4:8b > 92:65:f3:ec:b0:19, ethertype 802.1Q (0x8100), length 46: vlan 500, p 0, ethertype ARP (0x0806), Reply 192.168.50.2 is-at e4:b9:7a:f7:c4:8b, length 28 14:49:26.325091 92:65:f3:ec:b0:19 > e4:b9:7a:f7:c4:8b, ethertype 802.1Q (0x8100), length 102: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.1 > 192.168.50.2: ICMP echo request, id 44246, seq 1, length 64 14:49:26.325158 e4:b9:7a:f7:c4:8b > 92:65:f3:ec:b0:19, ethertype 802.1Q (0x8100), length 102: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2 > 192.168.50.1: ICMP echo reply, id 44246, seq 1, length 64on r3 i see these packets going out (so far it looks good):
root@bpi-r3:~# ping 192.168.60.2 PING 192.168.60.2 (192.168.60.2) 56(84) bytes of data. 13:30:29.782003 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 13:30:30.788175 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 13:30:31.828181 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 From 192.168.60.1 icmp_seq=1 Destination Host Unreachable 13:30:32.868205 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28 13:30:33.908171 08:22:33:44:55:77 > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 46: vlan 600, p 0, ethertype ARP (0x0806), Request who-has 192.168.60.2 tell 192.168.60.1, length 28
Quoted 1 line
HTH, maybe daniel or anyone other can confirm this
Does this help?
---
net/dsa/tag_mtk.c +8 -0
@@ -25,6 +25,14 @@ static struct sk_buff *mtk_tag_xmit(struu8 xmit_tpid;u8 *mtk_tag; +/* The Ethernet switch we are interfaced with needs packets to be at +* least 64 bytes (including FCS) otherwise their padding might be +* corrupted. With tags enabled, we need to make sure that packets are +* at least 68 bytes (including FCS and tag). +*/ +if (__skb_put_padto (skb,ETH_ZLEN +MTK_HDR_LEN , false)) +return NULL ; +/* Build the special tag after the MAC Source Address. If VLAN header * is present, it's required that VLAN header and special tag is * being combined. Only in this way we can allow the switch can parse
Aw: Re: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Hi
Quoted 12 lines
Gesendet: Freitag, 30. Dezember 2022 um 15:58 Uhr
Von: "Felix Fietkau" <[email protected]>Does this help?
---net/dsa/tag_mtk.c +1
-0@@ -25,6 +25,14 @@ static struct sk_buff *mtk_tag_xmit(struu8 xmit_tpid;u8 *mtk_tag; +/* The Ethernet switch we are interfaced with needs packets to be at
net/dsa/tag_mtk.c static struct sk_buff *mtk_tag_xmit(stru · line 38
Quoted 10 lines
net/dsa/tag_mtk.c +7
-0+* least 64 bytes (including FCS) otherwise their padding might be +* corrupted. With tags enabled, we need to make sure that packets are +* at least 68 bytes (including FCS and tag). +*/ +if (__skb_put_padto (skb,ETH_ZLEN +MTK_HDR_LEN , false)) +return NULL ; +/* Build the special tag after the MAC Source Address. If VLAN header * is present, it's required that VLAN header and special tag is * being combined. Only in this way we can allow the switch can parse
no, i verified my vlan-setup ist right by adding additional device (my r2) into the same 2 vlans. My Laptop can ping both vlans of R2, but on r3/mt7986 only vlan500 (on eth1) works, not 600 on wan-port of mt7986.
see arp going out on r3, but not received in the other side (laptop/r2)....
any debug i can add for this? ethtool does not show stats for vlans, for wan i see no CRC/drops
regards Frank
Aw: Re: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
In reply to Felix Fietkauseems only tx is affected on r3, as i see packets on the vlan from my laptop
tcpdump on R3 (e4:b9:7a:f7:c4:8b is mac from laptop):
13:47:05.265508 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 577: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2.59389 > 192.168.50.255.21027: UDP, length 531 13:47:05.265548 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 577: vlan 600, p 0, ethertype IPv4 (0x0800), 192.168.60.2.59389 > 192.168.60.255.21027: UDP, length 531
regards Frank
Felix Fietkau
maintainerRe: Aw: Re: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Quoted 8 lines · Frank Wunderlich
seems only tx is affected on r3, as i see packets on the vlan from my laptop
tcpdump on R3 (e4:b9:7a:f7:c4:8b is mac from laptop):
13:47:05.265508 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (
0x8100), length 577: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2.59389 > 192.168.50.255.21027: UDP, length 531 13:47:05.265548 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 577: vlan 600, p 0, ethertype IPv4 (0x0800), 192.168.60.2.59389 > 192.168.60.255.21027: UDP, length 531regards Frank
I don't have a setup to test 6.2 on my MT7986 board right now, but I did test latest OpenWrt with my changes and couldn't reproduce the issue there. I checked the diff between my tree and upstream and didn't find any relevant differences in mtk_eth_soc.c
Not sure what's going on or how to narrow it down further.
- Felix
Aw: Re: Re: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Hi,
update on my tests...something in my systemd-networkd breaks vlan on dsa-port.
if i boot with this disabled vlan added manually works on dsa-port and gmac too. so felix you can keep my tested-by.
if vlan is working and i activate networkd afterwards vlans are still working...i guess systemd puts the wan-interface into some kind of non-vlan-mode (vlan-filtering?).
this is my basic systemd-config:
/etc/systemd/network/15-wan.network
[Match]
Name=wan
[Network]
BindCarrier=eth0
#static setup
Address=192.168.0.19/24
Gateway=192.168.0.10
DNS=192.168.0.10
IPForward=yes
/etc/systemd/network/20-lanbr.netdev
[NetDev]
Name=lanbr0
Kind=bridge
[Bridge]
DefaultPVID=1
VLANFiltering=1/etc/systemd/network/22-lanbr.network
[Match]
Name=lanbr0
[Network]
BindCarrier=eth0
ConfigureWithoutCarrier=true
Address=192.168.1.1/24
Address=fd00:A::10/64
[DHCPServer]
PoolOffset=100
PoolSize=150
/etc/systemd/network/05-eth0.network
[Match]
Name=eth0
[Network]
DHCP=no
LinkLocalAddressing=no
ConfigureWithoutCarrier=true
/etc/systemd/network/21-lanbr-bind.network
[Match]
Name=lan0 lan1 lan2 lan3
[Network]
Bridge=lanbr0
/etc/systemd/network/10-wan.link
[Match]
OriginalName=wan
[Link]
Name=wan
#MACAddressPolicy=none
#MACAddress=08:22:33:44:55:77
regards Frank
Quoted 23 lines
Gesendet: Freitag, 30. Dezember 2022 um 17:13 Uhr
Von: "Felix Fietkau" <[email protected]>
An: "Frank Wunderlich" <[email protected]>
Cc: [email protected], "John Crispin" <[email protected]>, "Sean Wang" <[email protected]>, "Mark Lee" <[email protected]>, "Lorenzo Bianconi" <[email protected]>, "David S. Miller" <[email protected]>, "Eric Dumazet" <[email protected]>, "Jakub Kicinski" <[email protected]>, "Paolo Abeni" <[email protected]>, "Matthias Brugger" <[email protected]>, "Russell King" <[email protected]>, [email protected], [email protected], [email protected]
Betreff: Re: Aw: Re: Re: [PATCH net v3 4/5] net: ethernet:mtk_eth_soc: drop generic vlan rx offload, only use DSA untaggingQuoted 8 lines · Frank Wunderlich
seems only tx is affected on r3, as i see packets on the vlan from my laptop
tcpdump on R3 (e4:b9:7a:f7:c4:8b is mac from laptop):
13:47:05.265508 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (
0x8100), length 577: vlan 500, p 0, ethertype IPv4 (0x0800), 192.168.50.2.59389 > 192.168.50.255.21027: UDP, length 531 13:47:05.265548 e4:b9:7a:f7:c4:8b > ff:ff:ff:ff:ff:ff, ethertype 802.1Q (0x8100), length 577: vlan 600, p 0, ethertype IPv4 (0x0800), 192.168.60.2.59389 > 192.168.60.255.21027: UDP, length 531regards Frank
I don't have a setup to test 6.2 on my MT7986 board right now, but I did test latest OpenWrt with my changes and couldn't reproduce the issue there. I checked the diff between my tree and upstream and didn't find any relevant differences in
mtk_eth_soc.c
Not sure what's going on or how to narrow it down further.- Felix
Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
In reply to Felix FietkauQuoted 11 lines · Felix Fietkau
Through testing I found out that hardware vlan rx offload support seems to have some hardware issues. At least when using multiple MACs and when receiving tagged packets on the secondary MAC, the hardware can sometimes start to emit wrong tags on the first MAC as well.
In order to avoid such issues, drop the feature configuration and use the offload feature only for DSA hardware untagging on MT7621/MT7622 devices which only use one MAC.
Tested-by Frank Wunderlich <[email protected]>Signed-off-by Felix Fietkau <[email protected]>
You can add this to all patches on the series.
Tested on Unielec U7621-06 (MT7621AT) and Bananapi BPI-R2 (MT7623NI) on latest netnext with buildroot as initramfs. These are tested with my fix [0] applied on top.
VLAN on DSA master gmac0.
Works on MT7621 SoC with multi-chip module MT7530 switch and MT7623 SoC with standalone MT7530 switch.
VLAN on DSA master gmac0 and non-DSA gmac1.
Works on MT7621 SoC with multi-chip module MT7530 switch and MT7623 SoC with standalone MT7530 switch.
VLAN on DSA master gmac1.
Can’t test on MT7621 as an unrelated issue prevents from testing.
Define port@6 and gmac0, otherwise gmac1 DSA master receives malformed frames from port@5. This issue appears only on MT7621 SoC with multi-chip module MT7530 switch.
Works on MT7623 SoC with standalone MT7530 switch.
VLAN on DSA master gmac0 and DSA master gmac1.
Works on MT7621 SoC with multi-chip module MT7530 and MT7623 SoC with standalone MT7530 switch switch after compensating an unrelated issue.
When both MACs are DSA masters, ping from gmac1 DSA master first, otherwise frames received from user ports won’t reach to gmac1 DSA master. This issue appears on MT7621 SoC with multi-chip module MT7530 switch and MT7623 SoC with standalone MT7530 switch.
It'd be great if you could take a look at these issues.
Network configuration:
For DSA master gmac0/gmac1
ip l add link lan3 name lan3.50 type vlan id 50
ip a add 192.168.3.1/24 dev lan3.50
ip l set up lan3 && ip l set up lan3.50
For non-DSA gmac1
ip l del lan3.50
ip l add link eth1 name eth1.50 type vlan id 50
ip a add 192.168.3.1/24 dev eth1.50
ip l set up eth1 && ip l set up eth1.50
Other side
ip l add link enp9s0 name enp9s0.50 type vlan id 50 ip a add 192.168.3.2/24 dev enp9s0.50
ip l set up enp9s0 && ip l set up enp9s0.50
[0] https://lore.kernel.org/netdev/[email protected]/
Arınç
Aw: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Hi,
just to confirm, my issues were not caused by this series, it was there before and now fixed with [1]
at least one issue arinc is mention is fixed with [2]
so please apply
regards Frank
[1] https://patchwork.kernel.org/project/linux-mediatek/patch/[email protected]/
[2] https://patchwork.kernel.org/project/linux-mediatek/patch/[email protected]/
Re: Aw: Re: [PATCH net v3 4/5] net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging
Quoted 5 lines · Frank Wunderlich
Hi,
just to confirm, my issues were not caused by this series, it was there before and now fixed with [1]
at least one issue arinc is mention is fixed with [2]
Actually, [2] fixes another issue that I didn't mention here. But all the issues I mentioned here is unrelated to this patch series, and I already have some patches to submit that will fix them, so this series is in the clear.
Quoted 2 lines
so please apply
Agreed.
Arınç
Felix Fietkau
maintainerPATCH v3 5/5 net: ethernet: mtk_eth_soc: ppe: fix L2 offloading with DSA untagging offload enabled
In reply to Felix FietkauCheck for skb metadata in order to detect the case where the DSA header is not present.
--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 6 +++--- drivers/net/ethernet/mediatek/mtk_ppe.c | 5 ++++- 2 files changed, 7 insertions(+), 4 deletions(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +3 -3
@@ -2039,9 +2039,6 @@ static int mtk_poll_rx(struct napi_struct *napi, int budget,skb_checksum_none_assert (skb); skb->protocol =eth_type_trans (skb, netdev); -if (reason ==MTK_PPE_CPU_REASON_HIT_UNBIND_RATE_REACHED ) -mtk_ppe_check_skb (eth->ppe[0 ], skb, hash); -/* When using VLAN untagging in combination with DSA, the * hardware treats the MTK special tag as a VLAN and untags it. */
@@ -2054,6 +2051,9 @@ static int mtk_poll_rx(struct napi_struct *napi, int budget,skb_dst_set_noref (skb, ð->dsa_meta[port]->dst); } +if (reason ==MTK_PPE_CPU_REASON_HIT_UNBIND_RATE_REACHED ) +mtk_ppe_check_skb (eth->ppe[0 ], skb, hash); +skb_record_rx_queue (skb,0 );napi_gro_receive (napi, skb);
drivers/net/ethernet/mediatek/mtk_ppe.c +4 -1
@@ -8,6 +8,7 @@#include <linux/platform_device.h> #include <linux/if_ether.h> #include <linux/if_vlan.h> +#include <net/dst_metadata.h> #include <net/dsa.h> #include "mtk_eth_soc.h" #include "mtk_ppe.h"
@@ -700,7 +701,9 @@ void __mtk_ppe_check_skb(struct mtk_ppe *ppe, struct sk_buff *skb, u16 hash) skb->dev->dsa_ptr->tag_ops->proto != DSA_TAG_PROTO_MTK) goto out; - tag +=4 ; +if (!skb_metadata_dst (skb)) + tag +=4 ; +if (get_unaligned_be16 (tag) !=ETH_P_8021Q )break ;
Signature · 1 line
2.38.1