| Message ID | 20231107-gemini-largeframe-fix-v3-3-e3803c080b75@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Fix large frames in the Gemini ethernet driver | expand |
On Tue, Nov 07, 2023 at 10:54:28AM +0100, Linus Walleij wrote: > The Gemini ethernet controller provides hardware checksumming > for frames up to 1514 bytes including ethernet headers but not > FCS. > > If we start sending bigger frames (after first bumping up the MTU > on both interfaces sending and receiveing the frames), truncated s/receiveing/receiving/ > packets start to appear on the target such as in this tcpdump > resulting from ping -s 1474: > > 23:34:17.241983 14:d6:4d:a8:3c:4f (oui Unknown) > bc:ae:c5:6b:a8:3d (oui Unknown), > ethertype IPv4 (0x0800), length 1514: truncated-ip - 2 bytes missing! > (tos 0x0, ttl 64, id 32653, offset 0, flags [DF], proto ICMP (1), length 1502) > OpenWrt.lan > Fecusia: ICMP echo request, id 1672, seq 50, length 1482 > > If we bypass the hardware checksumming and provide a software > fallback, everything starts working fine up to the max TX MTU > of 2047 bytes, for example ping -s2000 192.168.1.2: > > 00:44:29.587598 bc:ae:c5:6b:a8:3d (oui Unknown) > 14:d6:4d:a8:3c:4f (oui Unknown), > ethertype IPv4 (0x0800), length 2042: > (tos 0x0, ttl 64, id 51828, offset 0, flags [none], proto ICMP (1), length 2028) > Fecusia > OpenWrt.lan: ICMP echo reply, id 1683, seq 4, length 2008 > > The bit enabling to bypass hardware checksum (or any of the > "TSS" bits) are undocumented in the hardware reference manual. > The entire hardware checksum unit appears undocumented. The > conclusion that we need to use the "bypass" bit was found by > trial-and-error. > > Since no hardware checksum will happen, we slot in a software > checksum fallback. > > Check for the condition where we need to compute checksum on the > skb with either hardware or software using == CHECKSUM_PARTIAL instead > of != CHECKSUM_NONE which is an incomplete check according to > <linux/skbuff.h>. > > We delete the code disabling the hardware checksum for large > MTU:s: this is suboptimal because it will disable hardware "MTUs" maybe? > checksumming also on small packets which the checksumming > engine can handle just fine, which is a waste of resources. > > On the D-Link DIR-685 router this fixes a bug on the conduit > interface to the RTL8366RB DSA switch: as the switch needs to add > space for its tag it increases the MTU on the conduit interface > to 1504 and that means that when the router sends packages > of 1500 bytes these get an extra 4 bytes of DSA tag and the > transfer fails because of the erroneous hardware checksumming, > affecting such basic functionality as the LuCI web interface. > > Suggested-by: Vladimir Oltean <olteanv@gmail.com> > Fixes: 4d5ae32f5e1e ("net: ethernet: Add a driver for Gemini gigabit ethernet") > Signed-off-by: Linus Walleij <linus.walleij@linaro.org> > --- > drivers/net/ethernet/cortina/gemini.c | 34 +++++++++++++++++++++++----------- > 1 file changed, 23 insertions(+), 11 deletions(-) > > diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c > index b21a94b4ab5c..78287cfcbf63 100644 > --- a/drivers/net/ethernet/cortina/gemini.c > +++ b/drivers/net/ethernet/cortina/gemini.c > @@ -1145,6 +1145,7 @@ static int gmac_map_tx_bufs(struct net_device *netdev, struct sk_buff *skb, > dma_addr_t mapping; > unsigned short mtu; > void *buffer; > + int ret; > > mtu = ETH_HLEN; > mtu += netdev->mtu; > @@ -1159,9 +1160,30 @@ static int gmac_map_tx_bufs(struct net_device *netdev, struct sk_buff *skb, > word3 |= mtu; > } > > - if (skb->ip_summed != CHECKSUM_NONE) { > + if (skb->len >= ETH_FRAME_LEN) { > + /* Hardware offloaded checksumming isn't working on frames > + * bigger than 1514 bytes. A hypothesis about this is that the > + * checksum buffer is only 1518 bytes, so when the frames get > + * bigger they get truncated, or the last few bytes get > + * overwritten by the FCS. > + * > + * Just use software checksumming and bypass on bigger frames. > + */ > + if (skb->ip_summed == CHECKSUM_PARTIAL) { > + ret = skb_checksum_help(skb); > + if (ret) > + return ret; > + } > + word1 |= TSS_BYPASS_BIT; > + } else if (skb->ip_summed == CHECKSUM_PARTIAL) { > int tcp = 0; > > + /* We do not switch off the checksumming on non TCP/UDP > + * frames: as is shown from tests, the checksumming engine > + * is smart enough to see that a frame is not actually TCP > + * or UDP and then just pass it through without any changes > + * to the frame. > + */ > if (skb->protocol == htons(ETH_P_IP)) { > word1 |= TSS_IP_CHKSUM_BIT; > tcp = ip_hdr(skb)->protocol == IPPROTO_TCP; > @@ -1978,15 +2000,6 @@ static int gmac_change_mtu(struct net_device *netdev, int new_mtu) > return 0; > } > > -static netdev_features_t gmac_fix_features(struct net_device *netdev, > - netdev_features_t features) > -{ > - if (netdev->mtu + ETH_HLEN + VLAN_HLEN > MTU_SIZE_BIT_MASK) > - features &= ~GMAC_OFFLOAD_FEATURES; > - > - return features; > -} > - I think this entire ndo_fix_features() can be indeed removed, but your justification was not immediately convincing. I'd point out that after your patch 1/4 "net: ethernet: cortina: Fix MTU max setting", you actually made this dead code, because netdev->mtu can't be larger than netdev->max_mtu. If you reverse the patch order a bit, such that "net: ethernet: cortina: Handle large frames" comes first, I think it would be much more logical for the removal of gmac_fix_features() to be part of the commit "net: ethernet: cortina: Fix MTU max setting", with the simple justification: the new MTU makes the code stop having any role. > static int gmac_set_features(struct net_device *netdev, > netdev_features_t features) > { > @@ -2212,7 +2225,6 @@ static const struct net_device_ops gmac_351x_ops = { > .ndo_set_mac_address = gmac_set_mac_address, > .ndo_get_stats64 = gmac_get_stats64, > .ndo_change_mtu = gmac_change_mtu, > - .ndo_fix_features = gmac_fix_features, > .ndo_set_features = gmac_set_features, > }; > > > -- > 2.34.1 >
diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c index b21a94b4ab5c..78287cfcbf63 100644 --- a/drivers/net/ethernet/cortina/gemini.c +++ b/drivers/net/ethernet/cortina/gemini.c @@ -1145,6 +1145,7 @@ static int gmac_map_tx_bufs(struct net_device *netdev, struct sk_buff *skb, dma_addr_t mapping; unsigned short mtu; void *buffer; + int ret; mtu = ETH_HLEN; mtu += netdev->mtu; @@ -1159,9 +1160,30 @@ static int gmac_map_tx_bufs(struct net_device *netdev, struct sk_buff *skb, word3 |= mtu; } - if (skb->ip_summed != CHECKSUM_NONE) { + if (skb->len >= ETH_FRAME_LEN) { + /* Hardware offloaded checksumming isn't working on frames + * bigger than 1514 bytes. A hypothesis about this is that the + * checksum buffer is only 1518 bytes, so when the frames get + * bigger they get truncated, or the last few bytes get + * overwritten by the FCS. + * + * Just use software checksumming and bypass on bigger frames. + */ + if (skb->ip_summed == CHECKSUM_PARTIAL) { + ret = skb_checksum_help(skb); + if (ret) + return ret; + } + word1 |= TSS_BYPASS_BIT; + } else if (skb->ip_summed == CHECKSUM_PARTIAL) { int tcp = 0; + /* We do not switch off the checksumming on non TCP/UDP + * frames: as is shown from tests, the checksumming engine + * is smart enough to see that a frame is not actually TCP + * or UDP and then just pass it through without any changes + * to the frame. + */ if (skb->protocol == htons(ETH_P_IP)) { word1 |= TSS_IP_CHKSUM_BIT; tcp = ip_hdr(skb)->protocol == IPPROTO_TCP; @@ -1978,15 +2000,6 @@ static int gmac_change_mtu(struct net_device *netdev, int new_mtu) return 0; } -static netdev_features_t gmac_fix_features(struct net_device *netdev, - netdev_features_t features) -{ - if (netdev->mtu + ETH_HLEN + VLAN_HLEN > MTU_SIZE_BIT_MASK) - features &= ~GMAC_OFFLOAD_FEATURES; - - return features; -} - static int gmac_set_features(struct net_device *netdev, netdev_features_t features) { @@ -2212,7 +2225,6 @@ static const struct net_device_ops gmac_351x_ops = { .ndo_set_mac_address = gmac_set_mac_address, .ndo_get_stats64 = gmac_get_stats64, .ndo_change_mtu = gmac_change_mtu, - .ndo_fix_features = gmac_fix_features, .ndo_set_features = gmac_set_features, };
The Gemini ethernet controller provides hardware checksumming for frames up to 1514 bytes including ethernet headers but not FCS. If we start sending bigger frames (after first bumping up the MTU on both interfaces sending and receiveing the frames), truncated packets start to appear on the target such as in this tcpdump resulting from ping -s 1474: 23:34:17.241983 14:d6:4d:a8:3c:4f (oui Unknown) > bc:ae:c5:6b:a8:3d (oui Unknown), ethertype IPv4 (0x0800), length 1514: truncated-ip - 2 bytes missing! (tos 0x0, ttl 64, id 32653, offset 0, flags [DF], proto ICMP (1), length 1502) OpenWrt.lan > Fecusia: ICMP echo request, id 1672, seq 50, length 1482 If we bypass the hardware checksumming and provide a software fallback, everything starts working fine up to the max TX MTU of 2047 bytes, for example ping -s2000 192.168.1.2: 00:44:29.587598 bc:ae:c5:6b:a8:3d (oui Unknown) > 14:d6:4d:a8:3c:4f (oui Unknown), ethertype IPv4 (0x0800), length 2042: (tos 0x0, ttl 64, id 51828, offset 0, flags [none], proto ICMP (1), length 2028) Fecusia > OpenWrt.lan: ICMP echo reply, id 1683, seq 4, length 2008 The bit enabling to bypass hardware checksum (or any of the "TSS" bits) are undocumented in the hardware reference manual. The entire hardware checksum unit appears undocumented. The conclusion that we need to use the "bypass" bit was found by trial-and-error. Since no hardware checksum will happen, we slot in a software checksum fallback. Check for the condition where we need to compute checksum on the skb with either hardware or software using == CHECKSUM_PARTIAL instead of != CHECKSUM_NONE which is an incomplete check according to <linux/skbuff.h>. We delete the code disabling the hardware checksum for large MTU:s: this is suboptimal because it will disable hardware checksumming also on small packets which the checksumming engine can handle just fine, which is a waste of resources. On the D-Link DIR-685 router this fixes a bug on the conduit interface to the RTL8366RB DSA switch: as the switch needs to add space for its tag it increases the MTU on the conduit interface to 1504 and that means that when the router sends packages of 1500 bytes these get an extra 4 bytes of DSA tag and the transfer fails because of the erroneous hardware checksumming, affecting such basic functionality as the LuCI web interface. Suggested-by: Vladimir Oltean <olteanv@gmail.com> Fixes: 4d5ae32f5e1e ("net: ethernet: Add a driver for Gemini gigabit ethernet") Signed-off-by: Linus Walleij <linus.walleij@linaro.org> --- drivers/net/ethernet/cortina/gemini.c | 34 +++++++++++++++++++++++----------- 1 file changed, 23 insertions(+), 11 deletions(-)