Re: [PATCHv2 net-next 5/5] selftests/net: add offload checking test for virtual interface

On Sat, Sep 06, 2025 at 11:30:58PM +0200, Sabrina Dubroca wrote:

quoted

+check_xfrm()
+{
+	local dev=$1
+	local src=192.0.2.1
+	local dst=192.0.2.2
+	local key="0x3132333435363738393031323334353664636261"
+
+	RET=0
+
+	ip -n "$ns" xfrm state flush
+	ip -n "$ns" xfrm state add proto esp src "$src" dst "$dst" spi 9 \
+		mode transport reqid 42 aead "rfc4106(gcm(aes))" "$key" 128 \
+		sel src "$src"/24 dst "$dst"/24 offload dev "$dev" dir out

It's maybe not something you would expect, but this codepath will not
check that NETIF_F_HW_ESP is set on $dev (you can verify that by
running "ip xfrm state add ... offload ..." on the same bond+netdevsim
combination before/after toggling esp-hw-offload on/off for the
bond). Why not use __check_offload again for this feature?

The esp-hw-offload is fixed on netdevsim

# ethtool -k eni0np1 | grep -i esp-hw-offload
esp-hw-offload: on [fixed]

There is no way to disable it. After we add the netdevsim to bond,
the bond also shows "esp-hw-offload off" as the flag is inherit
in dev->hw_enc_features, not dev->features.

It looks the only way to check if bond dev->hw_enc_features has NETIF_F_HW_ESP
is try set xfrm offload. As

static int xfrm_api_check(struct net_device *dev)
{
#ifdef CONFIG_XFRM_OFFLOAD
        if ((dev->features & NETIF_F_HW_ESP_TX_CSUM) &&
            !(dev->features & NETIF_F_HW_ESP))
                return NOTIFY_BAD;

        if ((dev->features & NETIF_F_HW_ESP) &&
            (!(dev->xfrmdev_ops &&
               dev->xfrmdev_ops->xdo_dev_state_add &&
               dev->xfrmdev_ops->xdo_dev_state_delete)))
                return NOTIFY_BAD;

Please correct me if I made any mistake.

Thanks
Hangbin