Skip to contentSkip to conversation

macsec: offload-related fixes

PATCH Superseded Sabrina Dubroca 3 participants 21 messages Oct 13, 2022 – Oct 25, 2022

Subsystems (3)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Sabrina Dubroca

author maintainer
#0 lore raw

cover macsec: offload-related fixes

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload. I'm working on re-adding those feature flags when offload is available, but I haven't fully solved that yet. I think it would be safer to do that second part in net-next considering how complex feature interactions tend to be.

Sabrina Dubroca (5):
Revert "net: macsec: report real_dev features when HW offloading is
enabled"
macsec: delete new rxsc when offload fails
macsec: fix secy->n_rx_sc accounting
macsec: fix detection of RXSCs when toggling offloading
macsec: clear encryption keys from the stack after setting up offload

 drivers/net/macsec.c | 51 ++++++++++++++++----------------------------
 1 file changed, 18 insertions(+), 33 deletions(-)
Signature · 1 line
2.38.0
author maintainer
#1 lore raw

PATCH 1/5 Revert "net: macsec: report real_dev features when HW offloading is enabled"

This reverts commit c850240b6c4132574a00f2da439277ab94265b66.

Copying the features of the real device doesn't work when it also supports IPsec offload, because then we'll have a device with NETIF_F_HW_ESP but no netdev->xfrmdev_ops.

Example perf trace when running
ip link add link eni1np1 type macsec port 4 offload mac

    ip   737 [003]   795.477676: probe:xfrm_dev_event__REGISTER      name="macsec0" features=0x1c000080014869
              xfrm_dev_event+0x3a
              notifier_call_chain+0x47
              register_netdevice+0x846
              macsec_newlink+0x25a

    ip   737 [003]   795.477687:   probe:xfrm_dev_event__return      ret=0x8002 (NOTIFY_BAD)
             notifier_call_chain+0x47
             register_netdevice+0x846
             macsec_newlink+0x25a

dev->features includes NETIF_F_HW_ESP (0x04000000000000), so xfrm_api_check returns NOTIFY_BAD because we don't have dev->xfrmdev_ops on the macsec device.

We could probably propagate GSO and a few other features from the lower device, similar to macvlan. This will be done in a future patch.

---
 drivers/net/macsec.c | 27 ++++-----------------------
 1 file changed, 4 insertions(+), 23 deletions(-)
drivers/net/macsec.c +4 -23
index c891b60937a7..b3f76e8071f2 100644
@@ -2654,11 +2654,6 @@ static int macsec_upd_offload(struct sk_buff *skb, struct genl_info *info)
 	if (ret)
 		goto rollback;
 
-	/* Force features update, since they are different for SW MACSec and
-	 * HW offloading cases.
-	 */
-	netdev_update_features(dev);
-
 	rtnl_unlock();
 	return 0;
 
@@ -3432,16 +3427,9 @@ static netdev_tx_t macsec_start_xmit(struct sk_buff *skb,
 	return ret;
 }
 
-#define SW_MACSEC_FEATURES \
+#define MACSEC_FEATURES \
 	(NETIF_F_SG | NETIF_F_HIGHDMA | NETIF_F_FRAGLIST)
 
-/* If h/w offloading is enabled, use real device features save for
- *   VLAN_FEATURES - they require additional ops
- *   HW_MACSEC - no reason to report it
- */
-#define REAL_DEV_FEATURES(dev) \
-	((dev)->features & ~(NETIF_F_VLAN_FEATURES | NETIF_F_HW_MACSEC))
-
 static int macsec_dev_init(struct net_device *dev)
 {
 	struct macsec_dev *macsec = macsec_priv(dev);
@@ -3458,12 +3446,8 @@ static int macsec_dev_init(struct net_device *dev)
 		return err;
 	}
 
-	if (macsec_is_offloaded(macsec)) {
-		dev->features = REAL_DEV_FEATURES(real_dev);
-	} else {
-		dev->features = real_dev->features & SW_MACSEC_FEATURES;
-		dev->features |= NETIF_F_LLTX | NETIF_F_GSO_SOFTWARE;
-	}
+	dev->features = real_dev->features & MACSEC_FEATURES;
+	dev->features |= NETIF_F_LLTX | NETIF_F_GSO_SOFTWARE;
 
 	dev->needed_headroom = real_dev->needed_headroom +
 			       MACSEC_NEEDED_HEADROOM;
@@ -3495,10 +3479,7 @@ static netdev_features_t macsec_fix_features(struct net_device *dev,
 	struct macsec_dev *macsec = macsec_priv(dev);
 	struct net_device *real_dev = macsec->real_dev;
 
-	if (macsec_is_offloaded(macsec))
-		return REAL_DEV_FEATURES(real_dev);
-
-	features &= (real_dev->features & SW_MACSEC_FEATURES) |
+	features &= (real_dev->features & MACSEC_FEATURES) |
 		    NETIF_F_GSO_SOFTWARE | NETIF_F_SOFT_FEATURES;
 	features |= NETIF_F_LLTX;
 
Signature · 1 line
2.38.0
author maintainer
#7 lore raw

PATCH 4/5 macsec: fix detection of RXSCs when toggling offloading

In reply to Sabrina Dubroca

macsec_is_configured incorrectly uses secy->n_rx_sc to check if some RXSCs exist. secy->n_rx_sc only counts the number of active RXSCs, but there can also be inactive SCs as well, which may be stored in the driver (in case we're disabling offloading), or would have to be pushed to the device (in case we're trying to enable offloading).

As long as RXSCs active on creation and never turned off, the issue is not visible.

Fixes dcb780fb2795 ("net: macsec: add nla support for changing the offloading selection")
---
 drivers/net/macsec.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/macsec.c +1 -1
index cdee342e42cd..0ad6eba0a807 100644
@@ -2572,7 +2572,7 @@ static bool macsec_is_configured(struct macsec_dev *macsec)
 	struct macsec_tx_sc *tx_sc = &secy->tx_sc;
 	int i;
 
-	if (secy->n_rx_sc > 0)
+	if (secy->rx_sc)
 		return true;
 
 	for (i = 0; i < MACSEC_NUM_AN; i++)
Signature · 1 line
2.38.0
author maintainer
#8 lore raw

PATCH 5/5 macsec: clear encryption keys from the stack after setting up offload

In reply to Sabrina Dubroca

macsec_add_rxsa and macsec_add_txsa copy the key to an on-stack offloading context to pass it to the drivers, but leaves it there when it's done. Clear it with memzero_explicit as soon as it's not needed anymore.

Fixes 3cf3227a21d1 ("net: macsec: hardware offloading infrastructure")
---
 drivers/net/macsec.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/net/macsec.c +2 -0
index 0ad6eba0a807..333e47b68e02 100644
@@ -1839,6 +1839,7 @@ static int macsec_add_rxsa(struct sk_buff *skb, struct genl_info *info)
 		       secy->key_len);
 
 		err = macsec_offload(ops->mdo_add_rxsa, &ctx);
+		memzero_explicit(ctx.sa.key, secy->key_len);
 		if (err)
 			goto cleanup;
 	}
@@ -2082,6 +2083,7 @@ static int macsec_add_txsa(struct sk_buff *skb, struct genl_info *info)
 		       secy->key_len);
 
 		err = macsec_offload(ops->mdo_add_txsa, &ctx);
+		memzero_explicit(ctx.sa.key, secy->key_len);
 		if (err)
 			goto cleanup;
 	}
Signature · 1 line
2.38.0
author maintainer
#2 lore raw

PATCH 2/5 macsec: delete new rxsc when offload fails

In reply to Sabrina Dubroca

Currently we get an inconsistent state:
- netlink returns the error to userspace
- the RXSC is installed but not offloaded

Then the device could get confused when we try to add an RXSA, because the RXSC isn't supposed to exist.

Fixes 3cf3227a21d1 ("net: macsec: hardware offloading infrastructure")
---
 drivers/net/macsec.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/net/macsec.c +2 -3
index b3f76e8071f2..0d6fe34b91ae 100644
@@ -1876,7 +1876,6 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	struct macsec_rx_sc *rx_sc;
 	struct nlattr *tb_rxsc[MACSEC_RXSC_ATTR_MAX + 1];
 	struct macsec_secy *secy;
-	bool was_active;
 	int ret;
 
 	if (!attrs[MACSEC_ATTR_IFINDEX])
@@ -1904,7 +1903,6 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 		return PTR_ERR(rx_sc);
 	}
 
-	was_active = rx_sc->active;
 	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
 		rx_sc->active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);
 
@@ -1931,7 +1929,8 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	return 0;
 
 cleanup:
-	rx_sc->active = was_active;
+	del_rx_sc(secy, sci);
+	free_rx_sc(rx_sc);
 	rtnl_unlock();
 	return ret;
 }
Signature · 1 line
2.38.0
author maintainer
#3 lore raw

PATCH 3/5 macsec: fix secy->n_rx_sc accounting

In reply to Sabrina Dubroca

secy->n_rx_sc is supposed to be the number of _active_ rxsc's within a secy. This is then used by macsec_send_sci to help decide if we should add the SCI to the header or not.

This logic is currently broken when we create a new RXSC and turn it off at creation, as create_rx_sc always sets ->active to true (and immediately uses that to increment n_rx_sc), and only later macsec_add_rxsc sets rx_sc->active.

Fixes c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver")
---
 drivers/net/macsec.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
drivers/net/macsec.c +9 -6
index 0d6fe34b91ae..cdee342e42cd 100644
@@ -1413,7 +1413,8 @@ static struct macsec_rx_sc *del_rx_sc(struct macsec_secy *secy, sci_t sci)
 	return NULL;
 }
 
-static struct macsec_rx_sc *create_rx_sc(struct net_device *dev, sci_t sci)
+static struct macsec_rx_sc *create_rx_sc(struct net_device *dev, sci_t sci,
+					 bool active)
 {
 	struct macsec_rx_sc *rx_sc;
 	struct macsec_dev *macsec;
@@ -1437,7 +1438,7 @@ static struct macsec_rx_sc *create_rx_sc(struct net_device *dev, sci_t sci)
 	}
 
 	rx_sc->sci = sci;
-	rx_sc->active = true;
+	rx_sc->active = active;
 	refcount_set(&rx_sc->refcnt, 1);
 
 	secy = &macsec_priv(dev)->secy;
@@ -1876,6 +1877,7 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	struct macsec_rx_sc *rx_sc;
 	struct nlattr *tb_rxsc[MACSEC_RXSC_ATTR_MAX + 1];
 	struct macsec_secy *secy;
+	bool active = true;
 	int ret;
 
 	if (!attrs[MACSEC_ATTR_IFINDEX])
@@ -1897,15 +1899,16 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	secy = &macsec_priv(dev)->secy;
 	sci = nla_get_sci(tb_rxsc[MACSEC_RXSC_ATTR_SCI]);
 
-	rx_sc = create_rx_sc(dev, sci);
+
+	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
+		active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);
+
+	rx_sc = create_rx_sc(dev, sci, active);
 	if (IS_ERR(rx_sc)) {
 		rtnl_unlock();
 		return PTR_ERR(rx_sc);
 	}
 
-	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
-		rx_sc->active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);
-
 	if (macsec_is_offloaded(netdev_priv(dev))) {
 		const struct macsec_ops *ops;
 		struct macsec_context ctx;
Signature · 1 line
2.38.0
#9 lore raw
In reply to Sabrina Dubroca
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

Quoted 2 lines

I'm working on re-adding those feature flags when offload is available, but I haven't fully solved that yet.

So let's revert when you are ready.

Thanks

#4 lore raw

Re: [PATCH net 3/5] macsec: fix secy->n_rx_sc accounting

In reply to Sabrina Dubroca
Quoted 47 lines · Sabrina Dubroca

secy->n_rx_sc is supposed to be the number of _active_ rxsc's within a secy. This is then used by macsec_send_sci to help decide if we should add the SCI to the header or not.

This logic is currently broken when we create a new RXSC and turn it off at creation, as create_rx_sc always sets ->active to true (and immediately uses that to increment n_rx_sc), and only later macsec_add_rxsc sets rx_sc->active.

Fixes c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver")
---
 drivers/net/macsec.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
drivers/net/macsec.c +4 -2
index 0d6fe34b91ae..cdee342e42cd 100644
@@ -1413,7 +1413,8 @@ static struct macsec_rx_sc *del_rx_sc(struct macsec_secy *secy, sci_t sci)
 	return NULL;
 }
 
-static struct macsec_rx_sc *create_rx_sc(struct net_device *dev, sci_t sci)
+static struct macsec_rx_sc *create_rx_sc(struct net_device *dev, sci_t sci,
+					 bool active)
 {
 	struct macsec_rx_sc *rx_sc;
 	struct macsec_dev *macsec;
@@ -1437,7 +1438,7 @@ static struct macsec_rx_sc *create_rx_sc(struct net_device *dev, sci_t sci)
 	}
 
 	rx_sc->sci = sci;
-	rx_sc->active = true;
+	rx_sc->active = active;
 	refcount_set(&rx_sc->refcnt, 1);
 
 	secy = &macsec_priv(dev)->secy;
@@ -1876,6 +1877,7 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	struct macsec_rx_sc *rx_sc;
 	struct nlattr *tb_rxsc[MACSEC_RXSC_ATTR_MAX + 1];
 	struct macsec_secy *secy;
+	bool active = true;
 	int ret;
 
 	if (!attrs[MACSEC_ATTR_IFINDEX])

drivers/net/macsec.c static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info) · line 1904

Quoted 7 lines
drivers/net/macsec.c +3 -1
@@ -1897,15 +1899,16 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	secy = &macsec_priv(dev)->secy;
 	sci = nla_get_sci(tb_rxsc[MACSEC_RXSC_ATTR_SCI]);
 
-	rx_sc = create_rx_sc(dev, sci);
+
+	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
+		active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);

You don't need !! to assign to bool variables and can safely omit them.

thanks

Quoted 16 lines
drivers/net/macsec.c +2 -3
+
+	rx_sc = create_rx_sc(dev, sci, active);
 	if (IS_ERR(rx_sc)) {
 		rtnl_unlock();
 		return PTR_ERR(rx_sc);
 	}
 
-	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
-		rx_sc->active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);
-
 	if (macsec_is_offloaded(netdev_priv(dev))) {
 		const struct macsec_ops *ops;
 		struct macsec_context ctx;
Signature · 1 line
2.38.0
author maintainer
#5 lore raw

Re: [PATCH net 3/5] macsec: fix secy->n_rx_sc accounting

2022-10-14, 09:16:56 +0300, Leon Romanovsky wrote:
[...]

Quoted 10 lines

drivers/net/macsec.c static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info) · line 1904

Quoted 8 lines
drivers/net/macsec.c +3 -1
@@ -1897,15 +1899,16 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	secy = &macsec_priv(dev)->secy;
 	sci = nla_get_sci(tb_rxsc[MACSEC_RXSC_ATTR_SCI]);
 
-	rx_sc = create_rx_sc(dev, sci);
+
+	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
+		active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);

You don't need !! to assign to bool variables and can safely omit them.

Yeah, but I'm just moving existing code, see below.

(please trim your replies, nobody wants to scroll through endless quoting just to find one comment hiding in the middle)

Quoted 12 lines

thanks

Quoted 9 lines
drivers/net/macsec.c +2 -2
+
+	rx_sc = create_rx_sc(dev, sci, active);
 	if (IS_ERR(rx_sc)) {
 		rtnl_unlock();
 		return PTR_ERR(rx_sc);
 	}
 
-	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
-		rx_sc->active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);
Signature · 1 line
Sabrina
author maintainer
#10 lore raw
In reply to Leon Romanovsky
Quoted 9 lines · Leon Romanovsky
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Quoted 5 lines

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

Yes, and that's not upstream yet. That patchset is also doing nothing to address the issue I'm refering to here, where xfrm_api_check rejects the macsec device because it has the NETIF_F_HW_ESP flag (passed from the lower device) and no xfrmdev_ops.

OTOH the mlx5 driver has both macsec and ipsec offload already, so I think it should be affected by this exact issue.

There are other feature flags that almost certainly shouldn't be copied from the lower device, such as the TLS offload flags.

Signature · 1 line
Sabrina
#11 lore raw
Quoted 10 lines · Sabrina Dubroca
Quoted 9 lines · Leon Romanovsky
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Quoted 7 lines
Quoted 5 lines

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

Yes, and that's not upstream yet.

For this conversation, you can assume what that series is merged. It is not merged due to request to change how we store XFRM policies in XFRM core code.

Quoted 3 lines

That patchset is also doing nothing to address the issue I'm refering to here, where xfrm_api_check rejects the macsec device because it has the NETIF_F_HW_ESP flag (passed from the lower device) and no xfrmdev_ops.

Of course, why do you think that IPsec series should address MACsec bugs?

Thanks

#6 lore raw

Re: [PATCH net 3/5] macsec: fix secy->n_rx_sc accounting

In reply to Sabrina Dubroca
Quoted 6 lines · Sabrina Dubroca

2022-10-14, 09:16:56 +0300, Leon Romanovsky wrote:
[...]

Quoted 4 lines
Quoted 4 lines
drivers/net/macsec.c +0 -0
@@ -1897,15 +1899,16 @@ static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info)
 	secy = &macsec_priv(dev)->secy;
 	sci = nla_get_sci(tb_rxsc[MACSEC_RXSC_ATTR_SCI]);
 
Quoted 8 lines
Quoted 6 lines

drivers/net/macsec.c static int macsec_add_rxsc(struct sk_buff *skb, struct genl_info *info) · line 1904

Quoted 4 lines
drivers/net/macsec.c +3 -1
-	rx_sc = create_rx_sc(dev, sci);
+
+	if (tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE])
+		active = !!nla_get_u8(tb_rxsc[MACSEC_RXSC_ATTR_ACTIVE]);

You don't need !! to assign to bool variables and can safely omit them.

Yeah, but I'm just moving existing code, see below.

Not really, original code was "rx_sc->active = ...", but you changed to use local value. So it is perfectly fine to fix the !! too.

Thanks

#12 lore raw
In reply to Leon Romanovsky

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 14 lines · Sabrina Dubroca
Quoted 13 lines
Quoted 12 lines · Leon Romanovsky
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

That patchset is also doing nothing to address the issue I'm refering to here, where xfrm_api_check rejects the macsec device because it has the NETIF_F_HW_ESP flag (passed from the lower device) and no xfrmdev_ops.

Of course, why do you think that IPsec series should address MACsec bugs?

I was looking at this and the series LGTM. I don't get the above concern, can you clarify?

If a lower device has both IPsec & MACsec offload capabilities:

- Without the revert: IPsec can be offloaded to the lower dev, MACsec can't. That's a bug.

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

Thanks,
Antoine

author maintainer
#20 lore raw
In reply to Leon Romanovsky
Quoted 23 lines · Leon Romanovsky
Quoted 10 lines · Sabrina Dubroca
Quoted 9 lines · Leon Romanovsky
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Quoted 7 lines
Quoted 5 lines

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

Yes, and that's not upstream yet.

For this conversation, you can assume what that series is merged. It is not merged due to request to change how we store XFRM policies in XFRM core code.

Quoted 5 lines
Quoted 3 lines

That patchset is also doing nothing to address the issue I'm refering to here, where xfrm_api_check rejects the macsec device because it has the NETIF_F_HW_ESP flag (passed from the lower device) and no xfrmdev_ops.

Of course, why do you think that IPsec series should address MACsec bugs?

I don't. That's why I really don't understand how the "full offload" series is relevant for this patchset.

Signature · 1 line
Sabrina
#13 lore raw
In reply to Antoine Tenart
Quoted 24 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 14 lines · Sabrina Dubroca
Quoted 13 lines
Quoted 12 lines · Leon Romanovsky
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

That patchset is also doing nothing to address the issue I'm refering to here, where xfrm_api_check rejects the macsec device because it has the NETIF_F_HW_ESP flag (passed from the lower device) and no xfrmdev_ops.

Of course, why do you think that IPsec series should address MACsec bugs?

Quoted 7 lines

I was looking at this and the series LGTM. I don't get the above concern, can you clarify?

If a lower device has both IPsec & MACsec offload capabilities:

- Without the revert: IPsec can be offloaded to the lower dev, MACsec can't. That's a bug.

And how does it possible that mlx5 macsec offload work?

Quoted 4 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

Thanks

Quoted 3 lines

Thanks,
Antoine

author maintainer
#14 lore raw
Quoted 28 lines · Leon Romanovsky
Quoted 27 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 14 lines · Sabrina Dubroca
Quoted 13 lines
Quoted 12 lines · Leon Romanovsky
Quoted 7 lines · Sabrina Dubroca

I'm working on a dummy offload for macsec on netdevsim. It just has a small SecY and RXSC table so I can trigger failures easily on the ndo_* side. It has exposed a couple of issues.

The first patch will cause some performance degradation, but in the current state it's not possible to offload macsec to lower devices that also support ipsec offload.

Please don't, IPsec offload is available and undergoing review.
https://lore.kernel.org/netdev/[email protected]/

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

This is whole series (XFRM + driver) for IPsec full offload.
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

That patchset is also doing nothing to address the issue I'm refering to here, where xfrm_api_check rejects the macsec device because it has the NETIF_F_HW_ESP flag (passed from the lower device) and no xfrmdev_ops.

Of course, why do you think that IPsec series should address MACsec bugs?

I was looking at this and the series LGTM. I don't get the above concern, can you clarify?

Quoted 6 lines
Quoted 4 lines

If a lower device has both IPsec & MACsec offload capabilities:

- Without the revert: IPsec can be offloaded to the lower dev, MACsec can't. That's a bug.

And how does it possible that mlx5 macsec offload work?

Well, I don't know. AFAICT, xfrm_api_check will be called for every new net_device created in the system, so a new macsec device with NETIF_F_HW_ESP will be rejected. Am I missing something?

I don't have access to mlx5 NICs with macsec offload so I can't check how that would work. My guess is that for some reason, the mlx5 driver didn't expose the NETIF_F_HW_ESP feature (CONFIG_MLX5_EN_IPSEC=n?). The only other possibility I see is CONFIG_XFRM=n.

Quoted 6 lines
Quoted 5 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

Quoted 6 lines

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

I don't understand what you're saying here. Patch #1 from this series is exactly about the macsec device acknowledging that ipsec offload exists. The rest of the patches is strictly macsec stuff and says nothing about ipsec. Can you point out where, in this series, I'm claiming that ipsec offload doesn't exist?

Signature · 1 line
Sabrina
#15 lore raw
Quoted 6 lines · Sabrina Dubroca
Quoted 5 lines · Leon Romanovsky
Quoted 4 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 3 lines · Sabrina Dubroca
Quoted 2 lines
Quoted 1 line · Leon Romanovsky

On Thu, Oct 13, 2022 at 04:15:38PM +0200, Sabrina Dubroca wrote:

<...>

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

Quoted 8 lines
Quoted 2 lines

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

I don't understand what you're saying here. Patch #1 from this series is exactly about the macsec device acknowledging that ipsec offload exists. The rest of the patches is strictly macsec stuff and says nothing about ipsec. Can you point out where, in this series, I'm claiming that ipsec offload doesn't exist?

All this conversation is about one sentence, which I cited above - "it's not possible to offload macsec to lower devices that also support ipsec offload". From the comments, I think that you wanted to say "macsec offload is not working due to performance optimization, where IPsec offload feature flag was exposed from lower device." Did I get it correctly, now?

Thanks

Quoted 4 lines
Signature · 2 lines
-- 
Sabrina
author maintainer
#16 lore raw
Quoted 27 lines · Leon Romanovsky
Quoted 6 lines · Sabrina Dubroca
Quoted 5 lines · Leon Romanovsky
Quoted 4 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 3 lines · Sabrina Dubroca
Quoted 2 lines
Quoted 1 line · Leon Romanovsky

On Thu, Oct 13, 2022 at 04:15:38PM +0200, Sabrina Dubroca wrote:

<...>

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

Quoted 8 lines
Quoted 2 lines

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

I don't understand what you're saying here. Patch #1 from this series is exactly about the macsec device acknowledging that ipsec offload exists. The rest of the patches is strictly macsec stuff and says nothing about ipsec. Can you point out where, in this series, I'm claiming that ipsec offload doesn't exist?

Quoted 5 lines

All this conversation is about one sentence, which I cited above - "it's not possible to offload macsec to lower devices that also support ipsec offload". From the comments, I think that you wanted to say "macsec offload is not working due to performance optimization, where IPsec offload feature flag was exposed from lower device." Did I get it correctly, now?

Yes. "In the current state" (that I wrote in front of the sentence you quoted) refers to the changes introduced by commit c850240b6c41. The details are present in the commit message for patch 1.

Do you object to the revert, if I rephrase the justification, and then re-add the features that make sense in net-next?

Signature · 1 line
Sabrina
#17 lore raw
Quoted 34 lines · Sabrina Dubroca
Quoted 27 lines · Leon Romanovsky
Quoted 6 lines · Sabrina Dubroca
Quoted 5 lines · Leon Romanovsky
Quoted 4 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 3 lines · Sabrina Dubroca
Quoted 2 lines
Quoted 1 line · Leon Romanovsky

On Thu, Oct 13, 2022 at 04:15:38PM +0200, Sabrina Dubroca wrote:

<...>

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

Quoted 8 lines
Quoted 2 lines

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

I don't understand what you're saying here. Patch #1 from this series is exactly about the macsec device acknowledging that ipsec offload exists. The rest of the patches is strictly macsec stuff and says nothing about ipsec. Can you point out where, in this series, I'm claiming that ipsec offload doesn't exist?

Quoted 5 lines

All this conversation is about one sentence, which I cited above - "it's not possible to offload macsec to lower devices that also support ipsec offload". From the comments, I think that you wanted to say "macsec offload is not working due to performance optimization, where IPsec offload feature flag was exposed from lower device." Did I get it correctly, now?

Quoted 6 lines

Yes. "In the current state" (that I wrote in front of the sentence you quoted) refers to the changes introduced by commit c850240b6c41. The details are present in the commit message for patch 1.

Do you object to the revert, if I rephrase the justification, and then re-add the features that make sense in net-next?

I don't have any objections.

Thanks

Quoted 4 lines
Signature · 2 lines
-- 
Sabrina
author maintainer
#18 lore raw
Quoted 35 lines · Leon Romanovsky
Quoted 34 lines · Sabrina Dubroca
Quoted 32 lines · Leon Romanovsky
Quoted 6 lines · Sabrina Dubroca
Quoted 5 lines · Leon Romanovsky
Quoted 4 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 3 lines · Sabrina Dubroca
Quoted 2 lines
Quoted 1 line · Leon Romanovsky

On Thu, Oct 13, 2022 at 04:15:38PM +0200, Sabrina Dubroca wrote:

<...>

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

Quoted 8 lines
Quoted 2 lines

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

I don't understand what you're saying here. Patch #1 from this series is exactly about the macsec device acknowledging that ipsec offload exists. The rest of the patches is strictly macsec stuff and says nothing about ipsec. Can you point out where, in this series, I'm claiming that ipsec offload doesn't exist?

All this conversation is about one sentence, which I cited above - "it's not possible to offload macsec to lower devices that also support ipsec offload". From the comments, I think that you wanted to say "macsec offload is not working due to performance optimization, where IPsec offload feature flag was exposed from lower device." Did I get it correctly, now?

Quoted 8 lines
Quoted 6 lines

Yes. "In the current state" (that I wrote in front of the sentence you quoted) refers to the changes introduced by commit c850240b6c41. The details are present in the commit message for patch 1.

Do you object to the revert, if I rephrase the justification, and then re-add the features that make sense in net-next?

I don't have any objections.

Would this be ok for the cover letter?

----
The first patch is a revert of commit c850240b6c41 ("net: macsec:
report real_dev features when HW offloading is enabled"). That commit tried to improve the performance of macsec offload by taking advantage of some of the NIC's features, but in doing so, broke macsec offload when the lower device supports both macsec and ipsec offload, as the ipsec offload feature flags were copied from the real device. Since the macsec device doesn't provide xdo_* ops, the XFRM core rejects the registration of the new macsec device in xfrm_api_check.

I'm working on re-adding those feature flags when offload is available, but I haven't fully solved that yet. I think it would be safer to do that second part in net-next considering how complex feature interactions tend to be.
----

Do you want something added to the commit message of the revert as well?

Thanks.

Signature · 1 line
Sabrina
#19 lore raw
Quoted 58 lines · Sabrina Dubroca
Quoted 35 lines · Leon Romanovsky
Quoted 34 lines · Sabrina Dubroca
Quoted 32 lines · Leon Romanovsky
Quoted 6 lines · Sabrina Dubroca
Quoted 5 lines · Leon Romanovsky
Quoted 4 lines · Antoine Tenart

Quoting Leon Romanovsky (2022-10-14 13:03:57)

Quoted 3 lines · Sabrina Dubroca
Quoted 2 lines
Quoted 1 line · Leon Romanovsky

On Thu, Oct 13, 2022 at 04:15:38PM +0200, Sabrina Dubroca wrote:

<...>

Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

- With the revert: IPsec and MACsec can be offloaded to the lower dev. Some features might not propagate to the MACsec dev, which won't allow some performance optimizations in the MACsec data path.

My concern is related to this sentence: "it's not possible to offload macsec to lower devices that also support ipsec offload", because our devices support both macsec and IPsec offloads at the same time.

Quoted 8 lines
Quoted 2 lines

I don't want to see anything (even in commit messages) that assumes that IPsec offload doesn't exist.

I don't understand what you're saying here. Patch #1 from this series is exactly about the macsec device acknowledging that ipsec offload exists. The rest of the patches is strictly macsec stuff and says nothing about ipsec. Can you point out where, in this series, I'm claiming that ipsec offload doesn't exist?

All this conversation is about one sentence, which I cited above - "it's not possible to offload macsec to lower devices that also support ipsec offload". From the comments, I think that you wanted to say "macsec offload is not working due to performance optimization, where IPsec offload feature flag was exposed from lower device." Did I get it correctly, now?

Quoted 8 lines
Quoted 6 lines

Yes. "In the current state" (that I wrote in front of the sentence you quoted) refers to the changes introduced by commit c850240b6c41. The details are present in the commit message for patch 1.

Do you object to the revert, if I rephrase the justification, and then re-add the features that make sense in net-next?

I don't have any objections.

Would this be ok for the cover letter?

----
The first patch is a revert of commit c850240b6c41 ("net: macsec:
report real_dev features when HW offloading is enabled"). That commit tried to improve the performance of macsec offload by taking advantage of some of the NIC's features, but in doing so, broke macsec offload when the lower device supports both macsec and ipsec offload, as the ipsec offload feature flags were copied from the real device. Since the macsec device doesn't provide xdo_* ops, the XFRM core rejects the registration of the new macsec device in xfrm_api_check.

Quoted 8 lines

I'm working on re-adding those feature flags when offload is available, but I haven't fully solved that yet. I think it would be safer to do that second part in net-next considering how complex feature interactions tend to be.
----

Do you want something added to the commit message of the revert as well?

It will be great to see this sentence "commit tried to improve ... device in xfrm_api_check." in that revert patch. It makes everything clearer.

Thanks