linux: mvebu: backport mvneta XDP support fixes

This patch backports additional fixes for XDP support in the mvneta driver. These
changes are found upstream as commits:

b37fa92e20ef2 net: mvneta: fix build skb for bm capable devices
f383b2950070c net: mvneta: rely on page_pool_recycle_direct in mvneta_run_xdp
79572c98c554d mvneta driver disallow XDP program on hardware buffer management
44efc78d0e464 net: mvneta: fix XDP support if sw bm is used as fallback

Signed-off-by: Jakov Petrina <jakov.petrina@sartura.hr>
master
Jakov Petrina 4 years ago committed by Luka Perkov
parent 76f9aa6f4b
commit 3dbb658d26

@ -0,0 +1,41 @@
From b37fa92e20ef28aada852cbf03b368d29a20478c Mon Sep 17 00:00:00 2001
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Thu, 14 Nov 2019 01:25:55 +0200
Subject: [PATCH] net: mvneta: fix build skb for bm capable devices
Fix build_skb for bm capable devices when they fall-back using swbm path
(e.g. when bm properties are configured in device tree but
CONFIG_MVNETA_BM_ENABLE is not set). In this case rx_offset_correction is
overwritten so we need to use it building skb instead of
MVNETA_SKB_HEADROOM directly
Fixes: 8dc9a0888f4c ("net: mvneta: rely on build_skb in mvneta_rx_swbm poll routine")
Fixes: 0db51da7a8e9 ("net: mvneta: add basic XDP support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reported-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/net/ethernet/marvell/mvneta.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -2165,7 +2165,7 @@ mvneta_swbm_rx_frame(struct mvneta_port
prefetch(data);
xdp->data_hard_start = data;
- xdp->data = data + MVNETA_SKB_HEADROOM + MVNETA_MH_SIZE;
+ xdp->data = data + pp->rx_offset_correction + MVNETA_MH_SIZE;
xdp->data_end = xdp->data + data_len;
xdp_set_data_meta_invalid(xdp);
@@ -2230,7 +2230,7 @@ mvneta_swbm_add_rx_fragment(struct mvnet
/* refill descriptor with new buffer later */
skb_add_rx_frag(rxq->skb,
skb_shinfo(rxq->skb)->nr_frags,
- page, MVNETA_SKB_HEADROOM, data_len,
+ page, pp->rx_offset_correction, data_len,
PAGE_SIZE);
}
page_pool_release_page(rxq->page_pool, page);

@ -0,0 +1,39 @@
From f383b2950070ce4f34e74db94f70bb565b746e97 Mon Sep 17 00:00:00 2001
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Wed, 20 Nov 2019 16:54:17 +0200
Subject: [PATCH] net: mvneta: rely on page_pool_recycle_direct in
mvneta_run_xdp
Rely on page_pool_recycle_direct and not on xdp_return_buff in
mvneta_run_xdp. This is a preliminary patch to limit the dma sync len
to the one strictly necessary
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/net/ethernet/marvell/mvneta.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -2108,7 +2108,8 @@ mvneta_run_xdp(struct mvneta_port *pp, s
err = xdp_do_redirect(pp->dev, xdp, prog);
if (err) {
ret = MVNETA_XDP_DROPPED;
- xdp_return_buff(xdp);
+ page_pool_recycle_direct(rxq->page_pool,
+ virt_to_head_page(xdp->data));
} else {
ret = MVNETA_XDP_REDIR;
}
@@ -2117,7 +2118,8 @@ mvneta_run_xdp(struct mvneta_port *pp, s
case XDP_TX:
ret = mvneta_xdp_xmit_back(pp, xdp);
if (ret != MVNETA_XDP_TX)
- xdp_return_buff(xdp);
+ page_pool_recycle_direct(rxq->page_pool,
+ virt_to_head_page(xdp->data));
break;
default:
bpf_warn_invalid_xdp_action(act);

@ -0,0 +1,53 @@
From 79572c98c554dcdb080bca547c871a51716dcdf8 Mon Sep 17 00:00:00 2001
From: Sven Auhagen <sven.auhagen@voleatech.de>
Date: Sat, 25 Jan 2020 08:07:03 +0000
Subject: [PATCH] mvneta driver disallow XDP program on hardware buffer
management
Recently XDP Support was added to the mvneta driver
for software buffer management only.
It is still possible to attach an XDP program if
hardware buffer management is used.
It is not doing anything at that point.
The patch disallows attaching XDP programs to mvneta
if hardware buffer management is used.
I am sorry about that. It is my first submission and I am having
some troubles with the format of my emails.
v4 -> v5:
- Remove extra tabs
v3 -> v4:
- Please ignore v3 I accidentally submitted
my other patch with git-send-mail and v4 is correct
v2 -> v3:
- My mailserver corrupted the patch
resubmission with git-send-email
v1 -> v2:
- Fixing the patches indentation
Signed-off-by: Sven Auhagen <sven.auhagen@voleatech.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/net/ethernet/marvell/mvneta.c | 6 ++++++
1 file changed, 6 insertions(+)
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -4231,6 +4231,12 @@ static int mvneta_xdp_setup(struct net_d
return -EOPNOTSUPP;
}
+ if (pp->bm_priv) {
+ NL_SET_ERR_MSG_MOD(extack,
+ "Hardware Buffer Management not supported on XDP");
+ return -EOPNOTSUPP;
+ }
+
need_update = !!pp->xdp_prog != !!prog;
if (running && need_update)
mvneta_stop(dev);

@ -0,0 +1,67 @@
From 44efc78d0e464ce70b45b165c005f8bedc17952e Mon Sep 17 00:00:00 2001
From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: Wed, 29 Jan 2020 12:50:53 +0100
Subject: [PATCH] net: mvneta: fix XDP support if sw bm is used as fallback
In order to fix XDP support if sw buffer management is used as fallback
for hw bm devices, define MVNETA_SKB_HEADROOM as maximum between
XDP_PACKET_HEADROOM and NET_SKB_PAD and let the hw aligns the IP header
to 4-byte boundary.
Fix rx_offset_correction initialization if mvneta_bm_port_init fails in
mvneta_resume routine
Fixes: 0db51da7a8e9 ("net: mvneta: add basic XDP support")
Tested-by: Sven Auhagen <sven.auhagen@voleatech.de>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/net/ethernet/marvell/mvneta.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -324,8 +324,7 @@
ETH_HLEN + ETH_FCS_LEN, \
cache_line_size())
-#define MVNETA_SKB_HEADROOM (max(XDP_PACKET_HEADROOM, NET_SKB_PAD) + \
- NET_IP_ALIGN)
+#define MVNETA_SKB_HEADROOM max(XDP_PACKET_HEADROOM, NET_SKB_PAD)
#define MVNETA_SKB_PAD (SKB_DATA_ALIGN(sizeof(struct skb_shared_info) + \
MVNETA_SKB_HEADROOM))
#define MVNETA_SKB_SIZE(len) (SKB_DATA_ALIGN(len) + MVNETA_SKB_PAD)
@@ -1172,6 +1171,7 @@ bm_mtu_err:
mvneta_bm_pool_destroy(pp->bm_priv, pp->pool_short, 1 << pp->id);
pp->bm_priv = NULL;
+ pp->rx_offset_correction = MVNETA_SKB_HEADROOM;
mvreg_write(pp, MVNETA_ACC_MODE, MVNETA_ACC_MODE_EXT1);
netdev_info(pp->dev, "fail to update MTU, fall back to software BM\n");
}
@@ -4954,7 +4954,6 @@ static int mvneta_probe(struct platform_
SET_NETDEV_DEV(dev, &pdev->dev);
pp->id = global_port_id++;
- pp->rx_offset_correction = MVNETA_SKB_HEADROOM;
/* Obtain access to BM resources if enabled and already initialized */
bm_node = of_parse_phandle(dn, "buffer-manager", 0);
@@ -4979,6 +4978,10 @@ static int mvneta_probe(struct platform_
}
of_node_put(bm_node);
+ /* sw buffer management */
+ if (!pp->bm_priv)
+ pp->rx_offset_correction = MVNETA_SKB_HEADROOM;
+
err = mvneta_init(&pdev->dev, pp);
if (err < 0)
goto err_netdev;
@@ -5136,6 +5139,7 @@ static int mvneta_resume(struct device *
err = mvneta_bm_port_init(pdev, pp);
if (err < 0) {
dev_info(&pdev->dev, "use SW buffer management\n");
+ pp->rx_offset_correction = MVNETA_SKB_HEADROOM;
pp->bm_priv = NULL;
}
}
Loading…
Cancel
Save