From: Simon Horman <horms@kernel.org>
To: Oleksij Rempel <o.rempel@pengutronix.de>
Cc: "David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
kernel@pengutronix.de, linux-kernel@vger.kernel.org,
netdev@vger.kernel.org,
Maxime Chevallier <maxime.chevallier@bootlin.com>
Subject: Re: [PATCH net-next v1 3/4] net: selftest: add checksum mode support and SW checksum handling
Date: Mon, 21 Apr 2025 15:32:46 +0100 [thread overview]
Message-ID: <20250421143246.GK2789685@horms.kernel.org> (raw)
In-Reply-To: <20250416161439.2922994-4-o.rempel@pengutronix.de>
On Wed, Apr 16, 2025 at 06:14:38PM +0200, Oleksij Rempel wrote:
> Introduce `enum net_test_checksum_mode` to support both CHECKSUM_COMPLETE
> and CHECKSUM_PARTIAL modes in selftest packet generation.
>
> Add helpers to calculate and apply software checksums for TCP/UDP in
> CHECKSUM_COMPLETE mode, and refactor checksum handling into a dedicated
> function `net_test_set_checksum()`.
>
> Update PHY loopback tests to use CHECKSUM_COMPLETE by default to avoid
> hardware offload dependencies and improve reliability.
>
> Also rename loopback test names to clarify checksum type and transport.
>
> Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Unfortunately this patch does not apply against net-next
(or at any rate, the series did not at the time it was submitted).
Please rebase and repost.
> ---
> net/core/selftests.c | 218 ++++++++++++++++++++++++++++++++++++++++---
> 1 file changed, 206 insertions(+), 12 deletions(-)
>
> diff --git a/net/core/selftests.c b/net/core/selftests.c
...
> +/**
> + * net_test_setup_sw_csum - Compute and apply software checksum
> + * (CHECKSUM_COMPLETE)
> + * @skb: Socket buffer with transport header set
> + * @iph: Pointer to IPv4 header inside skb
> + *
> + * This function computes and fills the transport layer checksum (TCP or UDP),
> + * and sets skb->ip_summed = CHECKSUM_COMPLETE.
> + *
> + * Returns 0 on success, or negative error code on failure.
> + */
> +static int net_test_setup_sw_csum(struct sk_buff *skb,
> + struct iphdr *iph)
> +{
> + int transport_offset = skb_transport_offset(skb);
> + int transport_len = skb->len - transport_offset;
> + __be16 final_csum;
> + __wsum csum;
> +
> + switch (iph->protocol) {
> + case IPPROTO_TCP:
> + if (!pskb_may_pull(skb,
> + transport_offset + sizeof(struct tcphdr)))
> + return -EFAULT;
> +
> + tcp_hdr(skb)->check = 0;
> + break;
> + case IPPROTO_UDP:
> + if (!pskb_may_pull(skb,
> + transport_offset + sizeof(struct udphdr)))
> + return -EFAULT;
> +
> + udp_hdr(skb)->check = 0;
> + break;
> + default:
> + pr_err("net_selftest: unsupported proto for sw csum: %u\n",
> + iph->protocol);
> + return -EINVAL;
> + }
> +
> + csum = skb_checksum(skb, transport_offset, transport_len, 0);
> + final_csum = csum_tcpudp_magic(iph->saddr, iph->daddr, transport_len,
> + iph->protocol, csum);
Sparse is unhappy about integer type annotations around here.
The 'final_csum =' line above is line number 101.
.../selftests.c:101:20: warning: incorrect type in assignment (different base types)
.../selftests.c:101:20: expected restricted __be16 [usertype] final_csum
.../selftests.c:101:20: got restricted __sum16
.../selftests.c:105:28: warning: incorrect type in assignment (different base types)
.../selftests.c:105:28: expected restricted __be16 [usertype] final_csum
.../selftests.c:105:28: got restricted __sum16 [usertype]
.../selftests.c:108:37: warning: incorrect type in assignment (different base types)
.../selftests.c:108:37: expected restricted __sum16 [usertype] check
.../selftests.c:108:37: got restricted __be16 [usertype] final_csum
.../selftests.c:110:37: warning: incorrect type in assignment (different base types)
.../selftests.c:110:37: expected restricted __sum16 [usertype] check
.../selftests.c:110:37: got restricted __be16 [usertype] final_csum
> +
> + if (iph->protocol == IPPROTO_UDP && final_csum == 0)
> + final_csum = CSUM_MANGLED_0;
> +
> + if (iph->protocol == IPPROTO_TCP)
> + tcp_hdr(skb)->check = final_csum;
> + else
> + udp_hdr(skb)->check = final_csum;
> +
> + skb->ip_summed = CHECKSUM_COMPLETE;
> +
> + return 0;
> +}
...
--
pw-bot: changes-requested
next prev parent reply other threads:[~2025-04-21 14:32 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-16 16:14 [PATCH net-next v1 0/4] net: selftest: improve test string formatting and checksum handling Oleksij Rempel
2025-04-16 16:14 ` [PATCH net-next v1 1/4] net: selftests: drop test index from net_selftest_get_strings() Oleksij Rempel
2025-04-21 14:33 ` Simon Horman
2025-04-16 16:14 ` [PATCH net-next v1 2/4] net: selftest: prepare for detailed error handling in net_test_get_skb() Oleksij Rempel
2025-04-21 14:33 ` Simon Horman
2025-04-16 16:14 ` [PATCH net-next v1 3/4] net: selftest: add checksum mode support and SW checksum handling Oleksij Rempel
2025-04-21 14:32 ` Simon Horman [this message]
2025-04-16 16:14 ` [PATCH net-next v1 4/4] net: selftest: add PHY loopback tests with HW checksum offload Oleksij Rempel
2025-04-21 14:33 ` Simon Horman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250421143246.GK2789685@horms.kernel.org \
--to=horms@kernel.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=kernel@pengutronix.de \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=maxime.chevallier@bootlin.com \
--cc=netdev@vger.kernel.org \
--cc=o.rempel@pengutronix.de \
--cc=pabeni@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).