public inbox for passt-dev@passt.top
 help / color / mirror / code / Atom feed
From: David Gibson <david@gibson.dropbear.id.au>
To: Laurent Vivier <lvivier@redhat.com>
Cc: passt-dev@passt.top
Subject: Re: [PATCH v8 11/30] tcp: Convert tcp_tap_handler() to use iov_tail
Date: Wed, 6 Aug 2025 12:35:14 +1000	[thread overview]
Message-ID: <aJK_Yp7ApUxKg2b0@zatzit> (raw)
In-Reply-To: <20250805154628.301343-12-lvivier@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 3465 bytes --]

On Tue, Aug 05, 2025 at 05:46:09PM +0200, Laurent Vivier wrote:
> Use packet_data() and extract headers using IOV_REMOVE_HEADER()
> and iov_remove_header_() rather than packet_get().
> 
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>

> ---
>  tcp.c | 31 ++++++++++++++++++++++++-------
>  1 file changed, 24 insertions(+), 7 deletions(-)
> 
> diff --git a/tcp.c b/tcp.c
> index 957b498db32d..f1048d7230c9 100644
> --- a/tcp.c
> +++ b/tcp.c
> @@ -310,6 +310,16 @@
>  #include "tcp_buf.h"
>  #include "tcp_vu.h"
>  
> +/*
> + * The size of TCP header (including options) is given by doff (Data Offset)
> + * that is a 4-bit value specifying the number of 32-bit words in the header.
> + * The maximum value of doff is 15 [(1 << 4) - 1].
> + * The maximum length in bytes of options is 15 minus the number of 32-bit
> + * words in the minimal TCP header (5) multiplied by the length of a 32-bit
> + * word (4).
> + */
> +#define OPTLEN_MAX (((1UL << 4) - 1 - 5) * 4UL)
> +
>  #ifndef __USE_MISC
>  /* From Linux UAPI, missing in netinet/tcp.h provided by musl */
>  struct tcp_repair_opt {
> @@ -1957,8 +1967,11 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af,
>  		    const struct pool *p, int idx, const struct timespec *now)
>  {
>  	struct tcp_tap_conn *conn;
> +	struct tcphdr th_storage;
>  	const struct tcphdr *th;
> -	size_t optlen, len;
> +	char optsc[OPTLEN_MAX];
> +	struct iov_tail data;
> +	size_t optlen, l4len;
>  	const char *opts;
>  	union flow *flow;
>  	flow_sidx_t sidx;
> @@ -1967,15 +1980,19 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af,
>  
>  	(void)pif;
>  
> -	th = packet_get(p, idx, 0, sizeof(*th), &len);
> +	if (!packet_data(p, idx, &data))
> +		return 1;
> +
> +	l4len = iov_tail_size(&data);
> +
> +	th = IOV_REMOVE_HEADER(&data, th_storage);
>  	if (!th)
>  		return 1;
> -	len += sizeof(*th);
>  
>  	optlen = th->doff * 4UL - sizeof(*th);
>  	/* Static checkers might fail to see this: */
> -	optlen = MIN(optlen, ((1UL << 4) /* from doff width */ - 6) * 4UL);
> -	opts = packet_get(p, idx, sizeof(*th), optlen, NULL);
> +	optlen = MIN(optlen, OPTLEN_MAX);

Pre-exisitng, but should we just drop packets with too many options,
rather than truncating the options?

> +	opts = (char *)iov_remove_header_(&data, &optsc[0], optlen, 1);
>  
>  	sidx = flow_lookup_af(c, IPPROTO_TCP, PIF_TAP, af, saddr, daddr,
>  			      ntohs(th->source), ntohs(th->dest));
> @@ -1987,7 +2004,7 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af,
>  			tcp_conn_from_tap(c, af, saddr, daddr, th,
>  					  opts, optlen, now);
>  		else
> -			tcp_rst_no_conn(c, af, saddr, daddr, flow_lbl, th, len);
> +			tcp_rst_no_conn(c, af, saddr, daddr, flow_lbl, th, l4len);
>  		return 1;
>  	}
>  
> @@ -1995,7 +2012,7 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af,
>  	ASSERT(pif_at_sidx(sidx) == PIF_TAP);
>  	conn = &flow->tcp;
>  
> -	flow_trace(conn, "packet length %zu from tap", len);
> +	flow_trace(conn, "packet length %zu from tap", l4len);
>  
>  	if (th->rst) {
>  		conn_event(c, conn, CLOSED);

-- 
David Gibson (he or they)	| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you, not the other way
				| around.
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

  reply	other threads:[~2025-08-06  3:06 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-08-05 15:45 [PATCH v8 00/30] Introduce discontiguous frames management Laurent Vivier
2025-08-05 15:45 ` [PATCH v8 01/30] arp: Don't mix incoming and outgoing buffers Laurent Vivier
2025-08-05 15:46 ` [PATCH v8 02/30] iov: Introduce iov_tail_clone() and iov_tail_drop() Laurent Vivier
2025-08-06  1:32   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 03/30] iov: Update IOV_REMOVE_HEADER() and IOV_PEEK_HEADER() Laurent Vivier
2025-08-06  1:45   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 04/30] tap: Use iov_tail with tap_add_packet() Laurent Vivier
2025-08-06  1:56   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 05/30] packet: Use iov_tail with packet_add() Laurent Vivier
2025-08-05 15:46 ` [PATCH v8 06/30] packet: Add packet_data() Laurent Vivier
2025-08-06  2:14   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 07/30] arp: Convert to iov_tail Laurent Vivier
2025-08-06  2:17   ` David Gibson
2025-08-07 12:58     ` Laurent Vivier
2025-08-07 13:11       ` Stefano Brivio
2025-08-13  2:21         ` David Gibson
2025-08-05 15:46 ` [PATCH v8 08/30] ndp: " Laurent Vivier
2025-08-05 15:46 ` [PATCH v8 09/30] icmp: " Laurent Vivier
2025-08-06  2:20   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 10/30] udp: " Laurent Vivier
2025-08-06  2:23   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 11/30] tcp: Convert tcp_tap_handler() to use iov_tail Laurent Vivier
2025-08-06  2:35   ` David Gibson [this message]
2025-08-05 15:46 ` [PATCH v8 12/30] tcp: Convert tcp_data_from_tap() " Laurent Vivier
2025-08-06  2:37   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 13/30] dhcpv6: move offset initialization out of dhcpv6_opt() Laurent Vivier
2025-08-05 15:46 ` [PATCH v8 14/30] dhcpv6: Extract sending of NotOnLink status Laurent Vivier
2025-08-05 15:46 ` [PATCH v8 15/30] dhcpv6: Convert to iov_tail Laurent Vivier
2025-08-05 15:46 ` [PATCH v8 16/30] dhcpv6: Use iov_tail in dhcpv6_opt() Laurent Vivier
2025-08-06  4:14   ` David Gibson
2025-08-08 13:59     ` Laurent Vivier
2025-08-13  2:29       ` David Gibson
2025-08-05 15:46 ` [PATCH v8 17/30] dhcp: Convert to iov_tail Laurent Vivier
2025-08-06  4:38   ` David Gibson
2025-08-08  9:33     ` Laurent Vivier
2025-08-13  2:27       ` David Gibson
2025-08-05 15:46 ` [PATCH v8 18/30] ip: Use iov_tail in ipv6_l4hdr() Laurent Vivier
2025-08-06  5:12   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 19/30] tap: Convert tap4_handler() to iov_tail Laurent Vivier
2025-08-06  5:17   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 20/30] tap: Convert tap6_handler() " Laurent Vivier
2025-08-06  6:21   ` David Gibson
2025-08-08 13:57     ` Laurent Vivier
2025-08-13  3:22       ` David Gibson
2025-08-05 15:46 ` [PATCH v8 21/30] packet: rename packet_data() to packet_get() Laurent Vivier
2025-08-06  6:22   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 22/30] arp: use iov_tail rather than pool Laurent Vivier
2025-08-06  6:24   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 23/30] dhcp: " Laurent Vivier
2025-08-06  6:26   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 24/30] dhcpv6: " Laurent Vivier
2025-08-06  6:27   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 25/30] icmp: " Laurent Vivier
2025-08-06  6:29   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 26/30] ndp: " Laurent Vivier
2025-08-06  6:31   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 27/30] packet: remove PACKET_POOL() and PACKET_POOL_P() Laurent Vivier
2025-08-06  6:32   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 28/30] packet: remove unused parameter from PACKET_POOL_DECL() Laurent Vivier
2025-08-06  6:33   ` David Gibson
2025-08-05 15:46 ` [PATCH v8 29/30] packet: Refactor vhost-user memory region handling Laurent Vivier
2025-08-07  6:10   ` David Gibson
2025-08-07  9:05     ` Laurent Vivier
2025-08-07 11:44       ` David Gibson
2025-08-05 15:46 ` [PATCH v8 30/30] packet: Add support for multi-vector packets Laurent Vivier
2025-08-07  6:17   ` David Gibson

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=aJK_Yp7ApUxKg2b0@zatzit \
    --to=david@gibson.dropbear.id.au \
    --cc=lvivier@redhat.com \
    --cc=passt-dev@passt.top \
    /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.
Code repositories for project(s) associated with this public inbox

	https://passt.top/passt

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 IMAP folder(s).