public inbox for passt-dev@passt.top
 help / color / mirror / code / Atom feed
From: Stefano Brivio <sbrivio@redhat.com>
To: David Gibson <david@gibson.dropbear.id.au>,
	Laurent Vivier <lvivier@redhat.com>
Cc: passt-dev@passt.top
Subject: Re: [PATCH 1/6] util: Add helper to find offset into io vector
Date: Tue, 27 Feb 2024 15:23:35 +0100	[thread overview]
Message-ID: <20240227152335.75ce9127@elisabeth> (raw)
In-Reply-To: <20240222055602.1872516-2-david@gibson.dropbear.id.au>

On Thu, 22 Feb 2024 16:55:57 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> tap_send_frames_passt() needs to determine which buffer element a byte
> offset into an IO vector lies in.  We have some upcoming uses for similar
> logic, so split this out into a helper function iov_offset().
> 
> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
>  tap.c  | 13 +++++--------
>  util.c | 23 +++++++++++++++++++++++
>  util.h |  1 +

Laurent, I guess this will need to be moved to iov.h by your series, at
the point where you introduce the new header.

>  3 files changed, 29 insertions(+), 8 deletions(-)
> 
> diff --git a/tap.c b/tap.c
> index 396dee7e..f15eba6e 100644
> --- a/tap.c
> +++ b/tap.c
> @@ -390,22 +390,19 @@ static size_t tap_send_frames_passt(const struct ctx *c,
>  		.msg_iovlen = n,
>  	};
>  	unsigned int i;
> +	size_t offset;
>  	ssize_t sent;
>  
>  	sent = sendmsg(c->fd_tap, &mh, MSG_NOSIGNAL | MSG_DONTWAIT);
>  	if (sent < 0)
>  		return 0;
>  
> -	/* Check for any partial frames due to short send */

Why would you drop this comment, though? I feel it's needed even more
as we don't open-code that any longer.

> -	for (i = 0; i < n; i++) {
> -		if ((size_t)sent < iov[i].iov_len)
> -			break;
> -		sent -= iov[i].iov_len;
> -	}
> +	offset = (size_t)sent;
> +	i = iov_offset(iov, n, &offset);

I think with these names and interface this becomes quite obscure: it
sounds like 'i' should be an offset at this point... and 'offset', I
have no idea (unless I read the comment to iov_offset()). Slightly
different proposal below.

>  
> -	if (i < n && sent) {
> +	if (i < n && offset) {
>  		/* A partial frame was sent */
> -		tap_send_remainder(c, &iov[i], sent);
> +		tap_send_remainder(c, &iov[i], offset);
>  		i++;
>  	}
>  
> diff --git a/util.c b/util.c
> index 21b35ff9..fb6a0430 100644
> --- a/util.c
> +++ b/util.c
> @@ -574,3 +574,26 @@ int do_clone(int (*fn)(void *), char *stack_area, size_t stack_size, int flags,
>  	return clone(fn, stack_area + stack_size / 2, flags, arg);
>  #endif
>  }
> +
> +/* iov_offset() - interpret offset into an IO vector
> + * @iov:	IO vector
> + * @n:		Number of entries in @iov
> + * @offset:	Pointer to offset into IO vector
> + *
> + * Return: index I of iovec which contains the given offset, or @n if
> + *         the given offset is >= the total # of bytes in the vector.
> + *         *@offset is updated to be the byte offset into (@iov + I),
> + *         and is guaranteed to be less than @iov[I].iov_len
> + */
> +size_t iov_offset(const struct iovec *iov, size_t n, size_t *offset)

...what do you think of:

/* iov_entry_index() - Index and optionally entry offset, given global offset
 * @iov:		IO vector
 * @n:			Number of entries in @iov
 * @global_offset:	Global offset of byte in IO vector we're looking for
 * @entry_offset:	If not NULL, set on return: entry offset
 *
 * Return: index of IO vector entry for given byte offset, @n if not found
 *
 * Note: @entry_offset is guaranteed to be less than @iov[i].iov_len, where i is
 * the return value
 */

and tap_send_frames_passt() could (more) happily do:

	i = iov_entry_index(iov, n, sent, &entry_offset);
	if (i < n) {
		/* A partial frame was sent */
		tap_send_remainder(c, &iov[i], entry_offset);
		i++;
	}

> +{
> +	size_t i;
> +
> +	for (i = 0; i < n; i++) {
> +	if (*offset < iov[i].iov_len)

Indentation.

> +			break;
> +		*offset -= iov[i].iov_len;
> +	}
> +
> +	return i;
> +}
> diff --git a/util.h b/util.h
> index d2320f8c..62fad6fe 100644
> --- a/util.h
> +++ b/util.h
> @@ -229,6 +229,7 @@ void write_pidfile(int fd, pid_t pid);
>  int __daemon(int pidfile_fd, int devnull_fd);
>  int fls(unsigned long x);
>  int write_file(const char *path, const char *buf);
> +size_t iov_offset(const struct iovec *iov, size_t n, size_t *offset);
>  
>  /**
>   * mod_sub() - Modular arithmetic subtraction

-- 
Stefano


  reply	other threads:[~2024-02-27 14:24 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-02-22  5:55 [PATCH 0/6] Allow more use of iovecs in pcap and tap interfaces David Gibson
2024-02-22  5:55 ` [PATCH 1/6] util: Add helper to find offset into io vector David Gibson
2024-02-27 14:23   ` Stefano Brivio [this message]
2024-02-27 23:27     ` David Gibson
2024-02-22  5:55 ` [PATCH 2/6] pcap: Update pcap_frame() to take an iovec and offset David Gibson
2024-02-27 14:23   ` Stefano Brivio
2024-02-22  5:55 ` [PATCH 3/6] util: Add write_remainder() helper David Gibson
2024-02-27 14:25   ` Stefano Brivio
2024-02-28  0:44     ` David Gibson
2024-02-28  6:24       ` Stefano Brivio
2024-02-28  9:04         ` David Gibson
2024-02-28  9:22           ` Stefano Brivio
2024-02-22  5:56 ` [PATCH 4/6] pcap: Handle short writes in pcap_frame() David Gibson
2024-02-22  5:56 ` [PATCH 5/6] pcap: Allow pcap_frame() and pcap_multiple() to take multi-buffer frames David Gibson
2024-02-27 14:26   ` Stefano Brivio
2024-02-22  5:56 ` [PATCH 6/6] tap: Use write_remainder() in tap_send_frames_passt() 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=20240227152335.75ce9127@elisabeth \
    --to=sbrivio@redhat.com \
    --cc=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).