Skip to contentSkip to conversation

selftests: drv-net: Test XDP native support

PATCH Applied Mohsin Bashir 8 participants 21 messages Jul 19, 2025 – Jul 24, 2025

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

v6 cover selftests: drv-net: Test XDP native support

This patch series add tests to validate XDP native support for PASS, DROP, ABORT, and TX actions, as well as headroom and tailroom adjustment. For adjustment tests, validate support for both the extension and shrinking cases across various packet sizes and offset values.

The pass criteria for head/tail adjustment tests require that at-least one adjustment value works for at-least one packet size. This ensure that the variability in maximum supported head/tail adjustment offset across different drivers is being incorporated.

The results reported in this series are based on netdevsim. However, the series is tested against multiple other drivers including fbnic.

Note: The XDP support for fbnic will be added later.
---
Change-log:
- Force checksum computation in netdevsim xdp hook
- Update checksum when updating packet for head/tail adjustment cases
- Use 1 as the minimum value for the data growth while adjusting tail

V5: https://lore.kernel.org/netdev/[email protected]
V4: https://lore.kernel.org/netdev/[email protected]
V3: https://lore.kernel.org/netdev/[email protected]
V2: https://lore.kernel.org/netdev/[email protected]
V1: https://lore.kernel.org/netdev/[email protected]

Jakub Kicinski (1):
net: netdevsim: hook in XDP handling

Mohsin Bashir (4):
selftests: drv-net: Test XDP_PASS/DROP support
selftests: drv-net: Test XDP_TX support
selftests: drv-net: Test tail-adjustment support
selftests: drv-net: Test head-adjustment support

 drivers/net/netdevsim/netdev.c                |  21 +-
 tools/testing/selftests/drivers/net/Makefile  |   1 +
 tools/testing/selftests/drivers/net/xdp.py    | 656 ++++++++++++++++++
 .../selftests/net/lib/xdp_native.bpf.c        | 621 +++++++++++++++++
 4 files changed, 1298 insertions(+), 1 deletion(-)
 create mode 100755 tools/testing/selftests/drivers/net/xdp.py
 create mode 100644 tools/testing/selftests/net/lib/xdp_native.bpf.c
Signature · 1 line
2.47.1
#1 lore raw

PATCH v6 1/5 net: netdevsim: hook in XDP handling

From: Jakub Kicinski <[email protected]>

Add basic XDP support by hooking in do_xdp_generic(). This should be enough to validate most basic XDP tests.

---
 drivers/net/netdevsim/netdev.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)
drivers/net/netdevsim/netdev.c +20 -1
index 611e7f65291c..a7628f5c09af 100644
@@ -387,15 +387,34 @@ static int nsim_get_iflink(const struct net_device *dev)
 static int nsim_rcv(struct nsim_rq *rq, int budget)
 {
 	struct net_device *dev = rq->napi.dev;
+	struct bpf_prog *xdp_prog;
+	struct netdevsim *ns;
 	struct sk_buff *skb;
 	unsigned int skblen;
 	int i, ret;
 
+	ns = netdev_priv(dev);
+	xdp_prog = READ_ONCE(ns->xdp.prog);
+
 	for (i = 0; i < budget; i++) {
 		if (skb_queue_empty(&rq->skb_queue))
 			break;
 
 		skb = skb_dequeue(&rq->skb_queue);
+
+		if (xdp_prog) {
+			/* skb might be freed directly by XDP, save the len */
+			skblen = skb->len;
+
+			if (skb->ip_summed == CHECKSUM_PARTIAL)
+				skb_checksum_help(skb);
+			ret = do_xdp_generic(xdp_prog, &skb);
+			if (ret != XDP_PASS) {
+				dev_dstats_rx_add(dev, skblen);
+				continue;
+			}
+		}
+
 		/* skb might be discard at netif_receive_skb, save the len */
 		skblen = skb->len;
 		skb_mark_napi_id(skb, &rq->napi);
@@ -936,7 +955,7 @@ static void nsim_setup(struct net_device *dev)
 			    NETIF_F_TSO;
 	dev->pcpu_stat_type = NETDEV_PCPU_STAT_DSTATS;
 	dev->max_mtu = ETH_MAX_MTU;
-	dev->xdp_features = NETDEV_XDP_ACT_HW_OFFLOAD;
+	dev->xdp_features = NETDEV_XDP_ACT_BASIC | NETDEV_XDP_ACT_HW_OFFLOAD;
 }
 
 static int nsim_queue_init(struct netdevsim *ns)
Signature · 1 line
2.47.1
#2 lore raw

PATCH v6 2/5 selftests: drv-net: Test XDP_PASS/DROP support

In reply to Mohsin Bashir

Test XDP_PASS/DROP in single buffer and multi buffer mode when XDP native support is available.

./drivers/net/xdp.py
TAP version 13
1..4
ok 1 xdp.test_xdp_native_pass_sb
ok 2 xdp.test_xdp_native_pass_mb
ok 3 xdp.test_xdp_native_drop_sb
ok 4 xdp.test_xdp_native_drop_mb
\# Totals: pass:4 fail:0 xfail:0 xpass:0 skip:0 error:0

---
 tools/testing/selftests/drivers/net/Makefile  |   1 +
 tools/testing/selftests/drivers/net/xdp.py    | 303 ++++++++++++++++++
 .../selftests/net/lib/xdp_native.bpf.c        | 158 +++++++++
 3 files changed, 462 insertions(+)
 create mode 100755 tools/testing/selftests/drivers/net/xdp.py
 create mode 100644 tools/testing/selftests/net/lib/xdp_native.bpf.c
tools/testing/selftests/drivers/net/Makefile +1 -0
index 9bd84d6b542e..3556f3563e08 100644
@@ -22,6 +22,7 @@ TEST_PROGS := \
 	stats.py \
 	shaper.py \
 	hds.py \
+	xdp.py \
 # end of TEST_PROGS
 
 include ../../lib.mk
tools/testing/selftests/drivers/net/xdp.py new file +303 -0

303 lines not on this page · raw message

tools/testing/selftests/net/lib/xdp_native.bpf.c new file +158 -0

158 lines not on this page · raw message

Signature · 1 line
2.47.1
#3 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 22 lines · Mohsin Bashir

Test XDP_PASS/DROP in single buffer and multi buffer mode when XDP native support is available.

./drivers/net/xdp.py
TAP version 13
1..4
ok 1 xdp.test_xdp_native_pass_sb
ok 2 xdp.test_xdp_native_pass_mb
ok 3 xdp.test_xdp_native_drop_sb
ok 4 xdp.test_xdp_native_drop_mb
\# Totals: pass:4 fail:0 xfail:0 xpass:0 skip:0 error:0

---
 tools/testing/selftests/drivers/net/Makefile  |   1 +
 tools/testing/selftests/drivers/net/xdp.py    | 303 ++++++++++++++++++
 .../selftests/net/lib/xdp_native.bpf.c        | 158 +++++++++
 3 files changed, 462 insertions(+)
 create mode 100755 tools/testing/selftests/drivers/net/xdp.py
 create mode 100644 tools/testing/selftests/net/lib/xdp_native.bpf.c

...

tools/testing/selftests/net/lib/xdp_native.bpf.c line 62

Quoted 11 lines
tools/testing/selftests/net/lib/xdp_native.bpf.c new file +11 -0
+
+static struct udphdr *filter_udphdr(struct xdp_md *ctx, __u16 port)
+{
+	void *data_end = (void *)(long)ctx->data_end;
+	void *data = (void *)(long)ctx->data;
+	struct udphdr *udph = NULL;
+	struct ethhdr *eth = data;
+
+	if (data + sizeof(*eth) > data_end)
+		return NULL;
+

This check assumes that the packet headers reside in the linear part of the xdp_buff. However, this assumption does not hold across all drivers. For example, in mlx5, the linear part is empty when using multi-buffer mode with striding rq configuration. This causes all multi-buffer test cases to fail over mlx5.

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Related discussion and context can be found here:
https://github.com/xdp-project/xdp-tools/pull/409

Quoted 30 lines
tools/testing/selftests/net/lib/xdp_native.bpf.c new file +30 -0
+	if (eth->h_proto == bpf_htons(ETH_P_IP)) {
+		struct iphdr *iph = data + sizeof(*eth);
+
+		if (iph + 1 > (struct iphdr *)data_end ||
+		    iph->protocol != IPPROTO_UDP)
+			return NULL;
+
+		udph = (void *)eth + sizeof(*iph) + sizeof(*eth);
+	} else if (eth->h_proto  == bpf_htons(ETH_P_IPV6)) {
+		struct ipv6hdr *ipv6h = data + sizeof(*eth);
+
+		if (ipv6h + 1 > (struct ipv6hdr *)data_end ||
+		    ipv6h->nexthdr != IPPROTO_UDP)
+			return NULL;
+
+		udph = (void *)eth + sizeof(*ipv6h) + sizeof(*eth);
+	} else {
+		return NULL;
+	}
+
+	if (udph + 1 > (struct udphdr *)data_end)
+		return NULL;
+
+	if (udph->dest != bpf_htons(port))
+		return NULL;
+
+	record_stats(ctx, STATS_RX);
+
+	return udph;
+}
maintainer
#4 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 17 lines · On Mon, 21 Jul 2025 14:43:15 +0300 Nimrod Oren wrote:

tools/testing/selftests/net/lib/xdp_native.bpf.c line 62

Quoted 10 lines
tools/testing/selftests/net/lib/xdp_native.bpf.c new file +10 -0
+static struct udphdr *filter_udphdr(struct xdp_md *ctx, __u16 port)
+{
+	void *data_end = (void *)(long)ctx->data_end;
+	void *data = (void *)(long)ctx->data;
+	struct udphdr *udph = NULL;
+	struct ethhdr *eth = data;
+
+	if (data + sizeof(*eth) > data_end)
+		return NULL;
+  

This check assumes that the packet headers reside in the linear part of the xdp_buff. However, this assumption does not hold across all drivers. For example, in mlx5, the linear part is empty when using multi-buffer mode with striding rq configuration. This causes all multi-buffer test cases to fail over mlx5.

Quoted 6 lines

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Related discussion and context can be found here:
https://github.com/xdp-project/xdp-tools/pull/409

That's a reasonable way to modify the test. But I'm not sure it's something that should be blocking merging the patches.
Or for that matter whether it's Mohsin's responsibility to make the test cater to quirks of mlx5, which is not even part of NIPA testing - we have no way of knowing what passes for mlx5, what regresses it etc.

Not related to this series, but I'd be curious what the perf hit is for having to load the headers.

#5 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 22 lines · Jakub Kicinski
Quoted 21 lines · On Mon, 21 Jul 2025 14:43:15 +0300 Nimrod Oren wrote:

tools/testing/selftests/net/lib/xdp_native.bpf.c line 62

Quoted 10 lines
tools/testing/selftests/net/lib/xdp_native.bpf.c new file +10 -0
+static struct udphdr *filter_udphdr(struct xdp_md *ctx, __u16 port)
+{
+	void *data_end = (void *)(long)ctx->data_end;
+	void *data = (void *)(long)ctx->data;
+	struct udphdr *udph = NULL;
+	struct ethhdr *eth = data;
+
+	if (data + sizeof(*eth) > data_end)
+		return NULL;
+  

This check assumes that the packet headers reside in the linear part of the xdp_buff. However, this assumption does not hold across all drivers. For example, in mlx5, the linear part is empty when using multi-buffer mode with striding rq configuration. This causes all multi-buffer test cases to fail over mlx5.

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Quoted 7 lines
Quoted 2 lines

Related discussion and context can be found here:
https://github.com/xdp-project/xdp-tools/pull/409

That's a reasonable way to modify the test. But I'm not sure it's something that should be blocking merging the patches. Or for that matter whether it's Mohsin's responsibility to make the test cater to quirks of mlx5,

Definitely not a quirk, you cannot assume the headers are in the linear part, especially if you're going to put this program as reference in the kernel tree.

This issue has nothing to do with mlx5, but a buggy XDP program.

Quoted 2 lines

which is not even part of NIPA testing -
we have no way of knowing what passes for mlx5, what regresses it etc.

People have been developing XDP code that runs on mlx5 long before NIPA even existed 🤷‍♂️..
And as you know, we run these selftests on mlx5 hardware, as evident by Nimrod's mail, and others you've seen on the list. You know what regresses.

maintainer
#6 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 10 lines · IOn Mon, 21 Jul 2025 21:34:05 +0300 Gal Pressman wrote:
Quoted 4 lines

That's a reasonable way to modify the test. But I'm not sure it's something that should be blocking merging the patches. Or for that matter whether it's Mohsin's responsibility to make the test cater to quirks of mlx5,

Definitely not a quirk, you cannot assume the headers are in the linear part, especially if you're going to put this program as reference in the kernel tree.

This issue has nothing to do with mlx5, but a buggy XDP program.

We put the tests in the tree to foster collaboration. If you think the test should be improved please send patches. I don't think the kernel will allow pulling headers if they are not in the linear section. But that's your problem to solve.

Quoted 7 lines
Quoted 2 lines

which is not even part of NIPA testing -
we have no way of knowing what passes for mlx5, what regresses it etc.

People have been developing XDP code that runs on mlx5 long before NIPA even existed 🤷‍♂️..
And as you know we run these selftests on mlx5 hardware, as evident by Nimrod's mail, and others you've seen on the list. You know what regresses.

No, please don't try to dispute facts. It's not integrated, if you go on a vacation upstream will have no idea what broke in mlx5. Either you are reporting the results upstream or our guarantees on regressions are best effort. BTW I don't understand how you can claim that a new test regresses something. It never passed on mlx5 == not a regression.

#7 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 10 lines · Jakub Kicinski
Quoted 9 lines · IOn Mon, 21 Jul 2025 21:34:05 +0300 Gal Pressman wrote:
Quoted 4 lines

That's a reasonable way to modify the test. But I'm not sure it's something that should be blocking merging the patches. Or for that matter whether it's Mohsin's responsibility to make the test cater to quirks of mlx5,

Definitely not a quirk, you cannot assume the headers are in the linear part, especially if you're going to put this program as reference in the kernel tree.

Quoted 6 lines
Quoted 1 line

This issue has nothing to do with mlx5, but a buggy XDP program.

We put the tests in the tree to foster collaboration. If you think the test should be improved please send patches. I don't think the kernel will allow pulling headers if they are not in the linear section. But that's your problem to solve.

Why send patches to fix something that shouldn't be merged broken in the first place?

Mohsin made an incorrect *assumption*, it is not based on anything that is guaranteed by XDP. Fixing this is trivial, and I don't see any technical reasons why not to.

Quoted 9 lines
Quoted 8 lines
Quoted 3 lines

which is not even part of NIPA testing -
we have no way of knowing what passes for mlx5, what regresses it etc.

People have been developing XDP code that runs on mlx5 long before NIPA even existed 🤷‍♂️..
And as you know we run these selftests on mlx5 hardware, as evident by Nimrod's mail, and others you've seen on the list. You know what regresses.

Quoted 5 lines

No, please don't try to dispute facts. It's not integrated, if you go on a vacation upstream will have no idea what broke in mlx5. Either you are reporting the results upstream or our guarantees on regressions are best effort. BTW I don't understand how you can claim that a new test regresses something. It never passed on mlx5 == not a regression.

This thread is more about pushing NIPA agenda than a technical discussion.

We're not part of NIPA as can be clearly observed by our MAINTAINERS entry. Don't confuse our NICs with others that are not accessible to the public, many people (inside and outside of NVIDIA) are developing and testing mlx5 without NIPA for years.

For the record, we run these selftests internally whether I go on vacation or not. These tests are running as part of our regression and failures are either reported to the list, or fixed by the team. You know all that.

maintainer
#8 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 1 line · On Tue, 22 Jul 2025 10:21:22 +0300 Gal Pressman wrote:

This thread is more about pushing NIPA agenda than a technical discussion.

What is the "NIPA agenda"? Please explain, this is getting funny :D

maintainer
#9 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

In reply to Gal Pressman
Quoted 31 lines · Gal Pressman
Quoted 22 lines · Jakub Kicinski
Quoted 21 lines · On Mon, 21 Jul 2025 14:43:15 +0300 Nimrod Oren wrote:

tools/testing/selftests/net/lib/xdp_native.bpf.c line 62

Quoted 10 lines
tools/testing/selftests/net/lib/xdp_native.bpf.c new file +10 -0
+static struct udphdr *filter_udphdr(struct xdp_md *ctx, __u16 port)
+{
+	void *data_end = (void *)(long)ctx->data_end;
+	void *data = (void *)(long)ctx->data;
+	struct udphdr *udph = NULL;
+	struct ethhdr *eth = data;
+
+	if (data + sizeof(*eth) > data_end)
+		return NULL;
+  

This check assumes that the packet headers reside in the linear part of the xdp_buff. However, this assumption does not hold across all drivers. For example, in mlx5, the linear part is empty when using multi-buffer mode with striding rq configuration. This causes all multi-buffer test cases to fail over mlx5.

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Quoted 7 lines
Quoted 2 lines

Related discussion and context can be found here:
https://github.com/xdp-project/xdp-tools/pull/409

That's a reasonable way to modify the test. But I'm not sure it's something that should be blocking merging the patches. Or for that matter whether it's Mohsin's responsibility to make the test cater to quirks of mlx5,

Quoted 5 lines

Definitely not a quirk, you cannot assume the headers are in the linear part, especially if you're going to put this program as reference in the kernel tree.

This issue has nothing to do with mlx5, but a buggy XDP program.

Note that with the self-tests we have a slightly different premise WRT the actual kernel code. We prefer on-boarding tests cases that work for some/most of the possible setup vs perfect ones, and eventually improve incrementally as needed: the goal is to increase the code coverage _and_ encourage people to contribute new tests upstream.

We try to avoid breaking existing tests (at least the ones actually reporting into the infrastructure), but for new ones the barriers are intentionally different than VS kernel code.

Cheers,

Paolo

#10 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 36 lines · Paolo Abeni
Quoted 35 lines · Gal Pressman
Quoted 22 lines · Jakub Kicinski
Quoted 21 lines · On Mon, 21 Jul 2025 14:43:15 +0300 Nimrod Oren wrote:

tools/testing/selftests/net/lib/xdp_native.bpf.c line 62

Quoted 10 lines
tools/testing/selftests/net/lib/xdp_native.bpf.c new file +10 -0
+static struct udphdr *filter_udphdr(struct xdp_md *ctx, __u16 port)
+{
+	void *data_end = (void *)(long)ctx->data_end;
+	void *data = (void *)(long)ctx->data;
+	struct udphdr *udph = NULL;
+	struct ethhdr *eth = data;
+
+	if (data + sizeof(*eth) > data_end)
+		return NULL;
+  

This check assumes that the packet headers reside in the linear part of the xdp_buff. However, this assumption does not hold across all drivers. For example, in mlx5, the linear part is empty when using multi-buffer mode with striding rq configuration. This causes all multi-buffer test cases to fail over mlx5.

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Quoted 7 lines
Quoted 2 lines

Related discussion and context can be found here:
https://github.com/xdp-project/xdp-tools/pull/409

That's a reasonable way to modify the test. But I'm not sure it's something that should be blocking merging the patches. Or for that matter whether it's Mohsin's responsibility to make the test cater to quirks of mlx5,

Definitely not a quirk, you cannot assume the headers are in the linear part, especially if you're going to put this program as reference in the kernel tree.

Quoted 7 lines
Quoted 1 line

This issue has nothing to do with mlx5, but a buggy XDP program.

Note that with the self-tests we have a slightly different premise WRT the actual kernel code. We prefer on-boarding tests cases that work for some/most of the possible setup vs perfect ones, and eventually improve incrementally as needed: the goal is to increase the code coverage _and_ encourage people to contribute new tests upstream.

The changes required to fix the test are trivial, and Nimrod provided a reference for the conversion.

#11 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

In reply to Nimrod Oren
Quoted 3 lines · Nimrod Oren

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Not sure if the test will be run on the older kernel. fwiw, there is a bpf_dynptr_slice_rdwr() kfunc which was added in commit 66e3a13e7c2c. It uses bpf_xdp_pointer() underneath, so it can avoid a copy if the requested data is contained in a single frag.

maintainer
#12 lore raw

Re: [PATCH net-next V6 2/5] selftests: drv-net: Test XDP_PASS/DROP support

Quoted 9 lines · On Mon, 21 Jul 2025 12:43:57 -0700 Martin KaFai Lau wrote:
Quoted 3 lines · Nimrod Oren

To ensure correctness across all drivers, all direct accesses to packet data should use these safer helper functions instead:
bpf_xdp_load_bytes() and bpf_xdp_store_bytes().

Not sure if the test will be run on the older kernel. fwiw, there is a bpf_dynptr_slice_rdwr() kfunc which was added in commit 66e3a13e7c2c. It uses bpf_xdp_pointer() underneath, so it can avoid a copy if the requested data is contained in a single frag.

We discussed running the tests on stable but nothing materialized, yet.

#13 lore raw

PATCH v6 3/5 selftests: drv-net: Test XDP_TX support

In reply to Mohsin Bashir

Add test to verify the XDP_TX functionality by generating traffic from a remote node on a specific UDP port and redirecting it back to the sender.

./drivers/net/xdp.py
TAP version 13
1..5
ok 1 xdp.test_xdp_native_pass_sb
ok 2 xdp.test_xdp_native_pass_mb
ok 3 xdp.test_xdp_native_drop_sb
ok 4 xdp.test_xdp_native_drop_mb
ok 5 xdp.test_xdp_native_tx_mb
\# Totals: pass:5 fail:0 xfail:0 xpass:0 skip:0 error:0

---
 tools/testing/selftests/drivers/net/xdp.py    | 34 ++++++++
 .../selftests/net/lib/xdp_native.bpf.c        | 80 +++++++++++++++++++
 2 files changed, 114 insertions(+)
tools/testing/selftests/drivers/net/xdp.py +34 -0
index 8ab1607edcd8..d42dbee78431 100755
@@ -27,6 +27,7 @@ class XDPAction(Enum):
     """Enum for XDP actions."""
     PASS = 0  # Pass the packet up to the stack
     DROP = 1  # Drop the packet
+    TX = 2    # Route the packet to the remote host
 
 
 class XDPStats(Enum):
@@ -34,6 +35,7 @@ class XDPStats(Enum):
     RX = 0    # Count of valid packets received for testing
     PASS = 1  # Count of packets passed up to the stack
     DROP = 2  # Count of packets dropped
+    TX = 3    # Count of incoming packets routed to the remote host
 
 
 @dataclass
@@ -180,6 +182,8 @@ def _get_stats(xdp_map_id):
             stats_formatted[XDPStats.PASS.value] = val
         elif stats_dump[key]["formatted"]["key"] == XDPStats.DROP.value:
             stats_formatted[XDPStats.DROP.value] = val
+        elif stats_dump[key]["formatted"]["key"] == XDPStats.TX.value:
+            stats_formatted[XDPStats.TX.value] = val
 
     return stats_formatted
 
@@ -278,6 +282,35 @@ def test_xdp_native_drop_mb(cfg):
     _test_drop(cfg, bpf_info, 8000)
 
 
+def test_xdp_native_tx_mb(cfg):
+    """
+    Tests the XDP_TX action for a multi-buff case.
+
+    Args:
+        cfg: Configuration object containing network settings.
+    """
+    cfg.require_cmd("socat", remote=True)
+
+    bpf_info = BPFProgInfo("xdp_prog_frags", "xdp_native.bpf.o", "xdp.frags", 9000)
+    prog_info = _load_xdp_prog(cfg, bpf_info)
+    port = rand_port()
+
+    _set_xdp_map("map_xdp_setup", TestConfig.MODE.value, XDPAction.TX.value)
+    _set_xdp_map("map_xdp_setup", TestConfig.PORT.value, port)
+
+    test_string = ''.join(random.choice(string.ascii_lowercase) for _ in range(8000))
+    rx_udp = f"socat -{cfg.addr_ipver} -T 2 -u UDP-RECV:{port},reuseport STDOUT"
+    tx_udp = f"echo {test_string} | socat -t 2 -u STDIN UDP:{cfg.baddr}:{port}"
+
+    with bkg(rx_udp, host=cfg.remote, exit_wait=True) as rnc:
+        cmd(tx_udp, host=cfg.remote, shell=True)
+
+    stats = _get_stats(prog_info['maps']['map_xdp_stats'])
+
+    ksft_eq(rnc.stdout.strip(), test_string, "UDP packet exchange failed")
+    ksft_eq(stats[XDPStats.TX.value], 1, "TX stats mismatch")
+
+
 def main():
     """
     Main function to execute the XDP tests.
@@ -294,6 +327,7 @@ def main():
                 test_xdp_native_pass_mb,
                 test_xdp_native_drop_sb,
                 test_xdp_native_drop_mb,
+                test_xdp_native_tx_mb,
             ],
             args=(cfg,))
     ksft_exit()
tools/testing/selftests/net/lib/xdp_native.bpf.c +80 -0

104 lines not on this page · raw message

Signature · 1 line
2.47.1
#14 lore raw

PATCH v6 4/5 selftests: drv-net: Test tail-adjustment support

In reply to Mohsin Bashir

Add test to validate support for the two cases of tail adjustment: 1) tail extension, and 2) tail shrinking across different frame sizes and offset values. For each of the two cases, test both the single and multi-buffer cases by choosing appropriate packet size.

The negative offset value result in growing of tailroom (shrinking of payload) while the positive offset result in shrinking of tailroom (growing of payload).

Since the support for tail adjustment varies across drivers, classify the test as pass if at least one combination of packet size and offset from a pre-selected list results in a successful run. In case of an unsuccessful run, report the failure and highlight the packet size and offset values that caused the test to fail, as well as the values that resulted in the last successful run.

Note: The growing part of this test for netdevsim may appear flaky when the offset value is larger than 1. This behavior occurs because tailroom is not explicitly reserved for netdevsim, with 1 being the typical tailroom value. However, in certain cases, such as payload being the last in the page with additional available space, the truesize is expanded. This also result increases the tailroom causing the test to pass intermittently. In contrast, when tailrrom is explicitly reserved, such as in the of fbnic, the test results are deterministic.

./drivers/net/xdp.py
TAP version 13
1..7
ok 1 xdp.test_xdp_native_pass_sb
ok 2 xdp.test_xdp_native_pass_mb
ok 3 xdp.test_xdp_native_drop_sb
ok 4 xdp.test_xdp_native_drop_mb
ok 5 xdp.test_xdp_native_tx_mb
\# Failed run: ... successful run: ... offset 1. Reason: Adjustment failed ok 6 xdp.test_xdp_native_adjst_tail_grow_data
ok 7 xdp.test_xdp_native_adjst_tail_shrnk_data
\# Totals: pass:7 fail:0 xfail:0 xpass:0 skip:0 error:0

---
 tools/testing/selftests/drivers/net/xdp.py    | 178 ++++++++++++++-
 .../selftests/net/lib/xdp_native.bpf.c        | 206 +++++++++++++++++-
 2 files changed, 380 insertions(+), 4 deletions(-)
tools/testing/selftests/drivers/net/xdp.py +176 -2
index d42dbee78431..85b5db0a1121 100755
@@ -11,7 +11,7 @@ import string
 from dataclasses import dataclass
 from enum import Enum
 
-from lib.py import ksft_run, ksft_exit, ksft_eq, ksft_ne
+from lib.py import ksft_run, ksft_exit, ksft_eq, ksft_ne, ksft_pr
 from lib.py import KsftFailEx, NetDrvEpEnv
 from lib.py import bkg, cmd, rand_port
 from lib.py import ip, bpftool, defer
@@ -21,6 +21,8 @@ class TestConfig(Enum):
     """Enum for XDP configuration options."""
     MODE = 0  # Configures the BPF program for a specific test
     PORT = 1  # Port configuration to communicate with the remote host
+    ADJST_OFFSET = 2  # Tail/Head adjustment offset for extension/shrinking
+    ADJST_TAG = 3  # Adjustment tag to annotate the start and end of extension
 
 
 class XDPAction(Enum):
@@ -28,6 +30,7 @@ class XDPAction(Enum):
     PASS = 0  # Pass the packet up to the stack
     DROP = 1  # Drop the packet
     TX = 2    # Route the packet to the remote host
+    TAIL_ADJST = 3  # Adjust the tail of the packet
 
 
 class XDPStats(Enum):
@@ -36,6 +39,7 @@ class XDPStats(Enum):
     PASS = 1  # Count of packets passed up to the stack
     DROP = 2  # Count of packets dropped
     TX = 3    # Count of incoming packets routed to the remote host
+    ABORT = 4 # Count of packets that were aborted
 
 
 @dataclass
@@ -174,7 +178,7 @@ def _get_stats(xdp_map_id):
         raise KsftFailEx(f"Failed to get stats for map {xdp_map_id}")
 
     stats_formatted = {}
-    for key in range(0, 4):
+    for key in range(0, 5):
         val = stats_dump[key]["formatted"]["value"]
         if stats_dump[key]["formatted"]["key"] == XDPStats.RX.value:
             stats_formatted[XDPStats.RX.value] = val
@@ -184,6 +188,8 @@ def _get_stats(xdp_map_id):
             stats_formatted[XDPStats.DROP.value] = val
         elif stats_dump[key]["formatted"]["key"] == XDPStats.TX.value:
             stats_formatted[XDPStats.TX.value] = val
+        elif stats_dump[key]["formatted"]["key"] == XDPStats.ABORT.value:
+            stats_formatted[XDPStats.ABORT.value] = val
 
     return stats_formatted
 
@@ -311,6 +317,172 @@ def test_xdp_native_tx_mb(cfg):
     ksft_eq(stats[XDPStats.TX.value], 1, "TX stats mismatch")
 
 
+def _validate_res(res, offset_lst, pkt_sz_lst):
+    """
+    Validates the result of a test.
+
+    Args:
+        res: The result of the test, which should be a dictionary with a "status" key.
+
+    Raises:
+        KsftFailEx: If the test fails to pass any combination of offset and packet size.
+    """
+    if "status" not in res:
+        raise KsftFailEx("Missing 'status' key in result dictionary")
+
+    # Validate that not a single case was successful
+    if res["status"] == "fail":
+        if res["offset"] == offset_lst[0] and res["pkt_sz"] == pkt_sz_lst[0]:
+            raise KsftFailEx(f"{res['reason']}")
+
+        # Get the previous offset and packet size to report the successful run
+        tmp_idx = offset_lst.index(res["offset"])
+        prev_offset = offset_lst[tmp_idx - 1]
+        if tmp_idx == 0:
+            tmp_idx = pkt_sz_lst.index(res["pkt_sz"])
+            prev_pkt_sz = pkt_sz_lst[tmp_idx - 1]
+        else:
+            prev_pkt_sz = res["pkt_sz"]
+
+        # Use these values for error reporting
+        ksft_pr(
+        f"Failed run: pkt_sz {res['pkt_sz']}, offset {res['offset']}. "
+        f"Last successful run: pkt_sz {prev_pkt_sz}, offset {prev_offset}. "
+        f"Reason: {res['reason']}"
+        )
+
+
+def _check_for_failures(recvd_str, stats):
+    """
+    Checks for common failures while adjusting headroom or tailroom.
+
+    Args:
+        recvd_str: The string received from the remote host after sending a test string.
+        stats: A dictionary containing formatted packet statistics for various XDP actions.
+
+    Returns:
+        str: A string describing the failure reason if a failure is detected, otherwise None.
+    """
+
+    # Any adjustment failure result in an abort hence, we track this counter
+    if stats[XDPStats.ABORT.value] != 0:
+        return "Adjustment failed"
+
+    # Since we are using aggregate stats for a single test across all offsets and packet sizes
+    # we can't use RX stats only to track data exchange failure without taking a previous
+    # snapshot. An easier way is to simply check for non-zero length of received string.
+    if len(recvd_str) == 0:
+        return "Data exchange failed"
+
+    # Check for RX and PASS stats mismatch. Ideally, they should be equal for a successful run
+    if stats[XDPStats.RX.value] != stats[XDPStats.PASS.value]:
+        return "RX stats mismatch"
+
+    return None
+
+
+def _test_xdp_native_tail_adjst(cfg, pkt_sz_lst, offset_lst):
+    """
+    Tests the XDP tail adjustment functionality.
+
+    This function loads the appropriate XDP program based on the provided
+    program name and configures the XDP map for tail adjustment. It then
+    validates the tail adjustment by sending and receiving UDP packets
+    with specified packet sizes and offsets.
+
+    Args:
+        cfg: Configuration object containing network settings.
+        prog: Name of the XDP program to load.
+        pkt_sz_lst: List of packet sizes to test.
+        offset_lst: List of offsets to validate support for tail adjustment.
+
+    Returns:
+        dict: A dictionary with test status and failure details if applicable.
+    """
+    port = rand_port()
+    bpf_info = BPFProgInfo("xdp_prog_frags", "xdp_native.bpf.o", "xdp.frags", 9000)
+
+    prog_info = _load_xdp_prog(cfg, bpf_info)
+
+    # Configure the XDP map for tail adjustment
+    _set_xdp_map("map_xdp_setup", TestConfig.MODE.value, XDPAction.TAIL_ADJST.value)
+    _set_xdp_map("map_xdp_setup", TestConfig.PORT.value, port)
+
+    for offset in offset_lst:
+        tag = format(random.randint(65, 90), "02x")
+
+        _set_xdp_map("map_xdp_setup", TestConfig.ADJST_OFFSET.value, offset)
+        if offset > 0:
+            _set_xdp_map("map_xdp_setup", TestConfig.ADJST_TAG.value, int(tag, 16))
+
+        for pkt_sz in pkt_sz_lst:
+            test_str = "".join(random.choice(string.ascii_lowercase) for _ in range(pkt_sz))
+            recvd_str = _exchg_udp(cfg, port, test_str)
+            stats = _get_stats(prog_info["maps"]["map_xdp_stats"])
+
+            failure = _check_for_failures(recvd_str, stats)
+            if failure is not None:
+                return {
+                    "status": "fail",
+                    "reason": failure,
+                    "offset": offset,
+                    "pkt_sz": pkt_sz,
+                }
+
+            # Validate data content based on offset direction
+            expected_data = None
+            if offset > 0:
+                expected_data = test_str + (offset * chr(int(tag, 16)))
+            else:
+                expected_data = test_str[0:pkt_sz + offset]
+
+            if recvd_str != expected_data:
+                return {
+                    "status": "fail",
+                    "reason": "Data mismatch",
+                    "offset": offset,
+                    "pkt_sz": pkt_sz,
+                }
+
+    return {"status": "pass"}
+
+
+def test_xdp_native_adjst_tail_grow_data(cfg):
+    """
+    Tests the XDP tail adjustment by growing packet data.
+
+    Args:
+        cfg: Configuration object containing network settings.
+    """
+    pkt_sz_lst = [512, 1024, 2048]
+    offset_lst = [1, 16, 32, 64, 128, 256]
+    res = _test_xdp_native_tail_adjst(
+        cfg,
+        pkt_sz_lst,
+        offset_lst,
+    )
+
+    _validate_res(res, offset_lst, pkt_sz_lst)
+
+
+def test_xdp_native_adjst_tail_shrnk_data(cfg):
+    """
+    Tests the XDP tail adjustment by shrinking packet data.
+
+    Args:
+        cfg: Configuration object containing network settings.
+    """
+    pkt_sz_lst = [512, 1024, 2048]
+    offset_lst = [-16, -32, -64, -128, -256]
+    res = _test_xdp_native_tail_adjst(
+        cfg,
+        pkt_sz_lst,
+        offset_lst,
+    )
+
+    _validate_res(res, offset_lst, pkt_sz_lst)
+
+
 def main():
     """
     Main function to execute the XDP tests.
@@ -328,6 +500,8 @@ def main():
                 test_xdp_native_drop_sb,
                 test_xdp_native_drop_mb,
                 test_xdp_native_tx_mb,
+                test_xdp_native_adjst_tail_grow_data,
+                test_xdp_native_adjst_tail_shrnk_data,
             ],
             args=(cfg,))
     ksft_exit()
tools/testing/selftests/net/lib/xdp_native.bpf.c +204 -2

253 lines not on this page · raw message

Signature · 1 line
2.47.1
#15 lore raw

PATCH v6 5/5 selftests: drv-net: Test head-adjustment support

In reply to Mohsin Bashir

Add test to validate the headroom adjustment support for both extension and the shrinking cases. For the extension part, eat up space from the start of payload data whereas, for the shrinking part, populate the newly available space with a tag. In the user-space, validate that a test string is manipulated accordingly.
The negative and positive offset values result in shrinking and growing of headroom (growing and shrinking of payload) respectively.

TAP version 13
1..9
ok 1 xdp.test_xdp_native_pass_sb
ok 2 xdp.test_xdp_native_pass_mb
ok 3 xdp.test_xdp_native_drop_sb
ok 4 xdp.test_xdp_native_drop_mb
ok 5 xdp.test_xdp_native_tx_mb
\# Failed run: pkt_sz 512, ... offset 1. Reason: Adjustment failed ok 6 xdp.test_xdp_native_adjst_tail_grow_data
ok 7 xdp.test_xdp_native_adjst_tail_shrnk_data
\# Failed run: pkt_sz 512, ... offset -128. Reason: Adjustment failed ok 8 xdp.test_xdp_native_adjst_head_grow_data
\# Failed run: pkt_sz (512) > HDS threshold (0) and offset 64 > 48 ok 9 xdp.test_xdp_native_adjst_head_shrnk_data
\# Totals: pass:9 fail:0 xfail:0 xpass:0 skip:0 error:0

---
 tools/testing/selftests/drivers/net/xdp.py    | 147 +++++++++++++-
 .../selftests/net/lib/xdp_native.bpf.c        | 181 ++++++++++++++++++
 2 files changed, 327 insertions(+), 1 deletion(-)
tools/testing/selftests/drivers/net/xdp.py +146 -1
index 85b5db0a1121..887d662ad128 100755
@@ -12,7 +12,7 @@ from dataclasses import dataclass
 from enum import Enum
 
 from lib.py import ksft_run, ksft_exit, ksft_eq, ksft_ne, ksft_pr
-from lib.py import KsftFailEx, NetDrvEpEnv
+from lib.py import KsftFailEx, NetDrvEpEnv, EthtoolFamily, NlError
 from lib.py import bkg, cmd, rand_port
 from lib.py import ip, bpftool, defer
 
@@ -31,6 +31,7 @@ class XDPAction(Enum):
     DROP = 1  # Drop the packet
     TX = 2    # Route the packet to the remote host
     TAIL_ADJST = 3  # Adjust the tail of the packet
+    HEAD_ADJST = 4  # Adjust the head of the packet
 
 
 class XDPStats(Enum):
@@ -483,6 +484,147 @@ def test_xdp_native_adjst_tail_shrnk_data(cfg):
     _validate_res(res, offset_lst, pkt_sz_lst)
 
 
+def get_hds_thresh(cfg):
+    """
+    Retrieves the header data split (HDS) threshold for a network interface.
+
+    Args:
+        cfg: Configuration object containing network settings.
+
+    Returns:
+        The HDS threshold value. If the threshold is not supported or an error occurs,
+        a default value of 1500 is returned.
+    """
+    netnl = cfg.netnl
+    hds_thresh = 1500
+
+    try:
+        rings = netnl.rings_get({'header': {'dev-index': cfg.ifindex}})
+        if 'hds-thresh' not in rings:
+            ksft_pr(f'hds-thresh not supported. Using default: {hds_thresh}')
+            return hds_thresh
+        hds_thresh = rings['hds-thresh']
+    except NlError as e:
+        ksft_pr(f"Failed to get rings: {e}. Using default: {hds_thresh}")
+
+    return hds_thresh
+
+
+def _test_xdp_native_head_adjst(cfg, prog, pkt_sz_lst, offset_lst):
+    """
+    Tests the XDP head adjustment action for a multi-buffer case.
+
+    Args:
+        cfg: Configuration object containing network settings.
+        netnl: Network namespace or link object (not used in this function).
+
+    This function sets up the packet size and offset lists, then performs
+    the head adjustment test by sending and receiving UDP packets.
+    """
+    cfg.require_cmd("socat", remote=True)
+
+    prog_info = _load_xdp_prog(cfg, BPFProgInfo(prog, "xdp_native.bpf.o", "xdp.frags", 9000))
+    port = rand_port()
+
+    _set_xdp_map("map_xdp_setup", TestConfig.MODE.value, XDPAction.HEAD_ADJST.value)
+    _set_xdp_map("map_xdp_setup", TestConfig.PORT.value, port)
+
+    hds_thresh = get_hds_thresh(cfg)
+    for offset in offset_lst:
+        for pkt_sz in pkt_sz_lst:
+            # The "head" buffer must contain at least the Ethernet header
+            # after we eat into it. We send large-enough packets, but if HDS
+            # is enabled head will only contain headers. Don't try to eat
+            # more than 28 bytes (UDPv4 + eth hdr left: (14 + 20 + 8) - 14)
+            l2_cut_off = 28 if cfg.addr_ipver == 4 else 48
+            if pkt_sz > hds_thresh and offset > l2_cut_off:
+                ksft_pr(
+                f"Failed run: pkt_sz ({pkt_sz}) > HDS threshold ({hds_thresh}) and "
+                f"offset {offset} > {l2_cut_off}"
+                )
+                return {"status": "pass"}
+
+            test_str = ''.join(random.choice(string.ascii_lowercase) for _ in range(pkt_sz))
+            tag = format(random.randint(65, 90), '02x')
+
+            _set_xdp_map("map_xdp_setup",
+                     TestConfig.ADJST_OFFSET.value,
+                     offset)
+            _set_xdp_map("map_xdp_setup", TestConfig.ADJST_TAG.value, int(tag, 16))
+            _set_xdp_map("map_xdp_setup", TestConfig.ADJST_OFFSET.value, offset)
+
+            recvd_str = _exchg_udp(cfg, port, test_str)
+
+            # Check for failures around adjustment and data exchange
+            failure = _check_for_failures(recvd_str, _get_stats(prog_info['maps']['map_xdp_stats']))
+            if failure is not None:
+                return {
+                    "status": "fail",
+                    "reason": failure,
+                    "offset": offset,
+                    "pkt_sz": pkt_sz
+                }
+
+            # Validate data content based on offset direction
+            expected_data = None
+            if offset < 0:
+                expected_data = chr(int(tag, 16)) * (0 - offset) + test_str
+            else:
+                expected_data = test_str[offset:]
+
+            if recvd_str != expected_data:
+                return {
+                    "status": "fail",
+                    "reason": "Data mismatch",
+                    "offset": offset,
+                    "pkt_sz": pkt_sz
+                }
+
+    return {"status": "pass"}
+
+
+def test_xdp_native_adjst_head_grow_data(cfg):
+    """
+    Tests the XDP headroom growth support.
+
+    Args:
+        cfg: Configuration object containing network settings.
+
+    This function sets up the packet size and offset lists, then calls the
+    _test_xdp_native_head_adjst_mb function to perform the actual test. The
+    test is passed if the headroom is successfully extended for given packet
+    sizes and offsets.
+    """
+    pkt_sz_lst = [512, 1024, 2048]
+
+    # Negative values result in headroom shrinking, resulting in growing of payload
+    offset_lst = [-16, -32, -64, -128, -256]
+    res = _test_xdp_native_head_adjst(cfg, "xdp_prog_frags", pkt_sz_lst, offset_lst)
+
+    _validate_res(res, offset_lst, pkt_sz_lst)
+
+
+def test_xdp_native_adjst_head_shrnk_data(cfg):
+    """
+    Tests the XDP headroom shrinking support.
+
+    Args:
+        cfg: Configuration object containing network settings.
+
+    This function sets up the packet size and offset lists, then calls the
+    _test_xdp_native_head_adjst_mb function to perform the actual test. The
+    test is passed if the headroom is successfully shrunk for given packet
+    sizes and offsets.
+    """
+    pkt_sz_lst = [512, 1024, 2048]
+
+    # Positive values result in headroom growing, resulting in shrinking of payload
+    offset_lst = [16, 32, 64, 128, 256]
+    res = _test_xdp_native_head_adjst(cfg, "xdp_prog_frags", pkt_sz_lst, offset_lst)
+
+    _validate_res(res, offset_lst, pkt_sz_lst)
+
+
 def main():
     """
     Main function to execute the XDP tests.
@@ -493,6 +635,7 @@ def main():
     function to execute the tests.
     """
     with NetDrvEpEnv(__file__) as cfg:
+        cfg.netnl = EthtoolFamily()
         ksft_run(
             [
                 test_xdp_native_pass_sb,
@@ -502,6 +645,8 @@ def main():
                 test_xdp_native_tx_mb,
                 test_xdp_native_adjst_tail_grow_data,
                 test_xdp_native_adjst_tail_shrnk_data,
+                test_xdp_native_adjst_head_grow_data,
+                test_xdp_native_adjst_head_shrnk_data,
             ],
             args=(cfg,))
     ksft_exit()
tools/testing/selftests/net/lib/xdp_native.bpf.c +181 -0

199 lines not on this page · raw message

Signature · 1 line
2.47.1
#16 lore raw
In reply to Mohsin Bashir
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Quoted 11 lines · On Sat, 19 Jul 2025 01:30:54 -0700 you wrote:

This patch series add tests to validate XDP native support for PASS, DROP, ABORT, and TX actions, as well as headroom and tailroom adjustment. For adjustment tests, validate support for both the extension and shrinking cases across various packet sizes and offset values.

The pass criteria for head/tail adjustment tests require that at-least one adjustment value works for at-least one packet size. This ensure that the variability in maximum supported head/tail adjustment offset across different drivers is being incorporated.

[...]

Here is the summary with links:
- [net-next,V6,1/5] net: netdevsim: hook in XDP handling

    https://git.kernel.org/netdev/net-next/c/be09f0d1acce
  - [net-next,V6,2/5] selftests: drv-net: Test XDP_PASS/DROP support
    https://git.kernel.org/netdev/net-next/c/1cbcb1b28b26
  - [net-next,V6,3/5] selftests: drv-net: Test XDP_TX support
    https://git.kernel.org/netdev/net-next/c/6713945726ce
  - [net-next,V6,4/5] selftests: drv-net: Test tail-adjustment support
    https://git.kernel.org/netdev/net-next/c/0b65cfcef9c5
  - [net-next,V6,5/5] selftests: drv-net: Test head-adjustment support
    https://git.kernel.org/netdev/net-next/c/d6444ebc97dc

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
#17 lore raw
Quoted 4 lines · On 23/07/2025 4:30, [email protected] wrote:

Hello:

This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Jakub, the test was still under active discussion, and you knowingly merged it with buggy code, ignoring our comments.

This is *extremely disrespectful* towards people that spent time reviewing, replying and participating in the netdev mailing list.

I expect better from you, and other netdev maintainers which allow this to happen. Please uphold your responsibilities as maintainers towards the community, this is unacceptable behavior.

maintainer
#18 lore raw
Quoted 9 lines · On Wed, 23 Jul 2025 10:11:57 +0300 Gal Pressman wrote:
Quoted 4 lines · On 23/07/2025 4:30, [email protected] wrote:

Hello:

This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Jakub, the test was still under active discussion, and you knowingly merged it with buggy code, ignoring our comments.

Quoted 6 lines

This is *extremely disrespectful* towards people that spent time reviewing, replying and participating in the netdev mailing list.

I expect better from you, and other netdev maintainers which allow this to happen. Please uphold your responsibilities as maintainers towards the community, this is unacceptable behavior.

Man, I just want tests upstream. How hard is this to understand.

These tests even work on mlx5, AFAIU, just not if LRO is enabled.

Please at least try to think about what's good for this community. At least try.

#19 lore raw
Quoted 13 lines · Jakub Kicinski
Quoted 12 lines · On Wed, 23 Jul 2025 10:11:57 +0300 Gal Pressman wrote:
Quoted 4 lines · On 23/07/2025 4:30, [email protected] wrote:

Hello:

This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Jakub, the test was still under active discussion, and you knowingly merged it with buggy code, ignoring our comments.

This is *extremely disrespectful* towards people that spent time reviewing, replying and participating in the netdev mailing list.

Quoted 7 lines
Quoted 3 lines

I expect better from you, and other netdev maintainers which allow this to happen. Please uphold your responsibilities as maintainers towards the community, this is unacceptable behavior.

Man, I just want tests upstream. How hard is this to understand.

These tests even work on mlx5, AFAIU, just not if LRO is enabled.

Incorrect.

Quoted 3 lines

Please at least try to think about what's good for this community. At least try.

Blatantly disrespecting the community is not good, it is unprofessional. Don't try to justify such actions.

reviewer
#20 lore raw
Quoted 22 lines · Gal Pressman
Quoted 13 lines · Jakub Kicinski
Quoted 12 lines · On Wed, 23 Jul 2025 10:11:57 +0300 Gal Pressman wrote:
Quoted 4 lines · On 23/07/2025 4:30, [email protected] wrote:

Hello:

This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Jakub, the test was still under active discussion, and you knowingly merged it with buggy code, ignoring our comments.

This is *extremely disrespectful* towards people that spent time reviewing, replying and participating in the netdev mailing list.

Quoted 7 lines
Quoted 3 lines

I expect better from you, and other netdev maintainers which allow this to happen. Please uphold your responsibilities as maintainers towards the community, this is unacceptable behavior.

Man, I just want tests upstream. How hard is this to understand.

These tests even work on mlx5, AFAIU, just not if LRO is enabled.

Quoted 8 lines

Incorrect.

Quoted 3 lines

Please at least try to think about what's good for this community. At least try.

Blatantly disrespecting the community is not good, it is unprofessional. Don't try to justify such actions.

All,

I'd like to ask everyone to take a step back and tone things down a couple of notches.

Let's work towards improving the tests. Rather than pursuing the current line of discussion.

Thanks!