[net,1/3] net: fec: dynamically set the NETDEV_XDP_ACT_NDO_XMIT feature of XDP

Message ID 20230704082916.2135501-2-wei.fang@nxp.com
State New
Headers
Series net: fec: fix some issues of ndo_xdp_xmit() |

Commit Message

Wei Fang July 4, 2023, 8:29 a.m. UTC
  From: Wei Fang <wei.fang@nxp.com>

When a XDP program is installed or uninstalled, fec_restart() will
be invoked to reset MAC and buffer descriptor rings. It's reasonable
not to transmit any packet during the process of reset. However, the
NETDEV_XDP_ACT_NDO_XMIT bit of xdp_features is enabled by default,
that is to say, it's possible that the fec_enet_xdp_xmit() will be
invoked even if the process of reset is not finished. In this case,
the redirected XDP frames might be dropped and available transmit BDs
may be incorrectly deemed insufficient. So this patch disable the
NETDEV_XDP_ACT_NDO_XMIT feature by default and dynamically configure
this feature when the bpf program is installed or uninstalled.

Fixes: e4ac7cc6e5a4 ("net: fec: turn on XDP features")
Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
 drivers/net/ethernet/freescale/fec_main.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
  

Comments

Andrew Lunn July 4, 2023, 11:41 p.m. UTC | #1
On Tue, Jul 04, 2023 at 04:29:14PM +0800, wei.fang@nxp.com wrote:
> From: Wei Fang <wei.fang@nxp.com>
> 
> When a XDP program is installed or uninstalled, fec_restart() will
> be invoked to reset MAC and buffer descriptor rings. It's reasonable
> not to transmit any packet during the process of reset. However, the
> NETDEV_XDP_ACT_NDO_XMIT bit of xdp_features is enabled by default,
> that is to say, it's possible that the fec_enet_xdp_xmit() will be
> invoked even if the process of reset is not finished. In this case,
> the redirected XDP frames might be dropped and available transmit BDs
> may be incorrectly deemed insufficient. So this patch disable the
> NETDEV_XDP_ACT_NDO_XMIT feature by default and dynamically configure
> this feature when the bpf program is installed or uninstalled.

I don't know much about XDP, so please excuse what might be a stupid
question.

Is this a generic issue? Should this
xdp_features_clear_redirect_target(dev) /
xdp_features_set_redirect_target(dev, false) be done in the core?

	Andrew
  
Toke Høiland-Jørgensen July 4, 2023, 11:54 p.m. UTC | #2
Andrew Lunn <andrew@lunn.ch> writes:

> On Tue, Jul 04, 2023 at 04:29:14PM +0800, wei.fang@nxp.com wrote:
>> From: Wei Fang <wei.fang@nxp.com>
>> 
>> When a XDP program is installed or uninstalled, fec_restart() will
>> be invoked to reset MAC and buffer descriptor rings. It's reasonable
>> not to transmit any packet during the process of reset. However, the
>> NETDEV_XDP_ACT_NDO_XMIT bit of xdp_features is enabled by default,
>> that is to say, it's possible that the fec_enet_xdp_xmit() will be
>> invoked even if the process of reset is not finished. In this case,
>> the redirected XDP frames might be dropped and available transmit BDs
>> may be incorrectly deemed insufficient. So this patch disable the
>> NETDEV_XDP_ACT_NDO_XMIT feature by default and dynamically configure
>> this feature when the bpf program is installed or uninstalled.
>
> I don't know much about XDP, so please excuse what might be a stupid
> question.
>
> Is this a generic issue? Should this
> xdp_features_clear_redirect_target(dev) /
> xdp_features_set_redirect_target(dev, false) be done in the core?

No, because not all drivers require an XDP program to be attached to
support being a redirect target (which is one of the reasons we
introduced these feature bits in the first place :)).

-Toke
  
Wei Fang July 5, 2023, 6:31 a.m. UTC | #3
> -----Original Message-----
> From: Toke Høiland-Jørgensen <toke@redhat.com>
> Sent: 2023年7月5日 7:54
> To: Andrew Lunn <andrew@lunn.ch>; Wei Fang <wei.fang@nxp.com>
> Cc: davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> pabeni@redhat.com; ast@kernel.org; daniel@iogearbox.net;
> hawk@kernel.org; john.fastabend@gmail.com; Shenwei Wang
> <shenwei.wang@nxp.com>; Clark Wang <xiaoning.wang@nxp.com>;
> netdev@vger.kernel.org; dl-linux-imx <linux-imx@nxp.com>;
> linux-kernel@vger.kernel.org; bpf@vger.kernel.org
> Subject: Re: [PATCH net 1/3] net: fec: dynamically set the
> NETDEV_XDP_ACT_NDO_XMIT feature of XDP
> 
> Andrew Lunn <andrew@lunn.ch> writes:
> 
> > On Tue, Jul 04, 2023 at 04:29:14PM +0800, wei.fang@nxp.com wrote:
> >> From: Wei Fang <wei.fang@nxp.com>
> >>
> >> When a XDP program is installed or uninstalled, fec_restart() will be
> >> invoked to reset MAC and buffer descriptor rings. It's reasonable not
> >> to transmit any packet during the process of reset. However, the
> >> NETDEV_XDP_ACT_NDO_XMIT bit of xdp_features is enabled by default,
> >> that is to say, it's possible that the fec_enet_xdp_xmit() will be
> >> invoked even if the process of reset is not finished. In this case,
> >> the redirected XDP frames might be dropped and available transmit BDs
> >> may be incorrectly deemed insufficient. So this patch disable the
> >> NETDEV_XDP_ACT_NDO_XMIT feature by default and dynamically
> configure
> >> this feature when the bpf program is installed or uninstalled.
> >
> > I don't know much about XDP, so please excuse what might be a stupid
> > question.
> >
> > Is this a generic issue? Should this
> > xdp_features_clear_redirect_target(dev) /
> > xdp_features_set_redirect_target(dev, false) be done in the core?
> 
> No, because not all drivers require an XDP program to be attached to support
> being a redirect target (which is one of the reasons we introduced these
> feature bits in the first place :)).
> 

Hi Toke,
Thanks for your explanation so much. :)
  

Patch

diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index 8fbe47703d47..9ce0319b33c3 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -3732,12 +3732,18 @@  static int fec_enet_bpf(struct net_device *dev, struct netdev_bpf *bpf)
 		if (fep->quirks & FEC_QUIRK_SWAP_FRAME)
 			return -EOPNOTSUPP;
 
+		if (!bpf->prog)
+			xdp_features_clear_redirect_target(dev);
+
 		if (is_run) {
 			napi_disable(&fep->napi);
 			netif_tx_disable(dev);
 		}
 
 		old_prog = xchg(&fep->xdp_prog, bpf->prog);
+		if (old_prog)
+			bpf_prog_put(old_prog);
+
 		fec_restart(dev);
 
 		if (is_run) {
@@ -3745,8 +3751,8 @@  static int fec_enet_bpf(struct net_device *dev, struct netdev_bpf *bpf)
 			netif_tx_start_all_queues(dev);
 		}
 
-		if (old_prog)
-			bpf_prog_put(old_prog);
+		if (bpf->prog)
+			xdp_features_set_redirect_target(dev, false);
 
 		return 0;
 
@@ -4016,8 +4022,7 @@  static int fec_enet_init(struct net_device *ndev)
 
 	if (!(fep->quirks & FEC_QUIRK_SWAP_FRAME))
 		ndev->xdp_features = NETDEV_XDP_ACT_BASIC |
-				     NETDEV_XDP_ACT_REDIRECT |
-				     NETDEV_XDP_ACT_NDO_XMIT;
+				     NETDEV_XDP_ACT_REDIRECT;
 
 	fec_restart(ndev);