All of lore.kernel.org
 help / color / mirror / Atom feed
From: Xiaowen Xia <haoyurenzhuxia@gmail.com>
To: Bagas Sanjaya <bagasdotme@gmail.com>
Cc: git@vger.kernel.org, gitster@pobox.com, peff@peff.net,
	worldhello.net@gmail.com,
	Xia XiaoWen <chenan.xxw@alibaba-inc.com>
Subject: Re: [PATCH] add http.maxReceiveSpeed to limit git-receive-pack receiving speed
Date: Thu, 19 Aug 2021 19:27:41 +0800	[thread overview]
Message-ID: <CAHLXgnZQmLiLbGnbC-nouE6tTJam5ZmdRmj=8E+tmM1pNLVFXw@mail.gmail.com> (raw)
In-Reply-To: <a097e447-28e7-d991-ac99-ce1ffdb3ead4@gmail.com>

> I mean "In some situations it is preferable to limit receiving speed of
> git clone/fetch due to network bandwidth constraint, otherwise other
> applications performance will suffer from degraded network performance.".
Thanks!

> Wouldn't it make sense for negative speed? Why don't just return error
> if negative speed is set?
You are right, negative speed doesn't make sense. but I prefer printf
a warning and continue work instead of returning an error, negative
speed seems not a fatal error.

> Why is there boilerplate #if #endif block for libcurl 7.15.5?
The speed limit depends on the `CURLOPT_MAX_RECV_SPEED_LARGE` provided
by libCURL, and libCURL added `CURLOPT_MAX_RECV_SPEED_LARGE` in
7.15.5: https://curl.se/changes.html#7_15_5 .

Bagas Sanjaya <bagasdotme@gmail.com> 于2021年8月19日周四 下午5:36写道:
>
> On 19/08/21 16.14, Xia XiaoWen wrote:
> > Sometimes need to limit the receive speed of git `clone/fetch`
> > because of the limited network bandwidth, otherwise will prevent
> > other applications from using the network normally.
> >
> I mean "In some situations it is preferable to limit receiving speed of
> git clone/fetch due to network bandwidth constraint, otherwise other
> applications performance will suffer from degraded network performance.".
>
> > Add `http.maxReceiveSpeed` to limit `git-receive-pack` receiving
> > speed, Can be overridden by `GIT_HTTP_MAX_RECEIVE_SPEED` eivironment
> > variable.
> >
> s/eivironment/environment/
>
> > The default is unlimited, same if the value is 0 or negative. The
> > default unit is Bytes/s, common unit suffixes of k, m, or g are supported.
> >
>
> Wouldn't it make sense for negative speed? Why don't just return error
> if negative speed is set?
>
> > this configuration is valid for `clone`, `fetch`, `pull` commands of the
> > https protocol, and only supports libcurl 7.15.5 and above.
> > ---
> >   http.c | 24 +++++++++++++++++++++++-
> >   1 file changed, 23 insertions(+), 1 deletion(-)
> >
> > diff --git a/http.c b/http.c
> > index 8119247149..12030cf3bc 100644
> > --- a/http.c
> > +++ b/http.c
> > @@ -83,6 +83,9 @@ static const char *ssl_pinnedkey;
> >   static const char *ssl_cainfo;
> >   static long curl_low_speed_limit = -1;
> >   static long curl_low_speed_time = -1;
> > +#if LIBCURL_VERSION_NUM >= 0x070f05
> > +static ssize_t curl_max_receive_speed = -1;
> > +#endif
> >   static int curl_ftp_no_epsv;
> >   static const char *curl_http_proxy;
> >   static const char *http_proxy_authmethod;
> > @@ -361,7 +364,12 @@ static int http_options(const char *var, const char *value, void *cb)
> >               curl_low_speed_time = (long)git_config_int(var, value);
> >               return 0;
> >       }
> > -
> > +#if LIBCURL_VERSION_NUM >= 0x070f05
> > +     if (!strcmp("http.maxreceivespeed", var)) {
> > +             curl_max_receive_speed = git_config_ssize_t(var, value);
> > +             return 0;
> > +     }
> > +#endif
> >       if (!strcmp("http.noepsv", var)) {
> >               curl_ftp_no_epsv = git_config_bool(var, value);
> >               return 0;
> > @@ -974,6 +982,12 @@ static CURL *get_curl_handle(void)
> >                                curl_low_speed_time);
> >       }
> >
> > +#if LIBCURL_VERSION_NUM >= 0x070f05
> > +     if (curl_max_receive_speed > 0)
> > +             curl_easy_setopt(result, CURLOPT_MAX_RECV_SPEED_LARGE,
> > +                              curl_max_receive_speed);
> > +#endif
> > +
> >       curl_easy_setopt(result, CURLOPT_MAXREDIRS, 20);
> >   #if LIBCURL_VERSION_NUM >= 0x071301
> >       curl_easy_setopt(result, CURLOPT_POSTREDIR, CURL_REDIR_POST_ALL);
> > @@ -1105,6 +1119,9 @@ void http_init(struct remote *remote, const char *url, int proactive_auth)
> >   {
> >       char *low_speed_limit;
> >       char *low_speed_time;
> > +#if LIBCURL_VERSION_NUM >= 0x070f05
> > +     char *max_receive_speed;
> > +#endif
> >       char *normalized_url;
> >       struct urlmatch_config config = { STRING_LIST_INIT_DUP };
> >
> > @@ -1196,6 +1213,11 @@ void http_init(struct remote *remote, const char *url, int proactive_auth)
> >       low_speed_time = getenv("GIT_HTTP_LOW_SPEED_TIME");
> >       if (low_speed_time != NULL)
> >               curl_low_speed_time = strtol(low_speed_time, NULL, 10);
> > +#if LIBCURL_VERSION_NUM >= 0x070f05
> > +     max_receive_speed = getenv("GIT_HTTP_MAX_RECEIVE_SPEED");
> > +     if (max_receive_speed && !git_parse_ssize_t(max_receive_speed, &curl_max_receive_speed))
> > +             warning("failed to parse GIT_HTTP_MAX_RECEIVE_SPEED: %s", max_receive_speed);
> > +#endif
> >
> >       if (curl_ssl_verify == -1)
> >               curl_ssl_verify = 1;
> >
>
> Why is there boilerplate #if #endif block for libcurl 7.15.5?
>
> --
> An old man doll... just what I always wanted! - Clara

  reply	other threads:[~2021-08-19 11:27 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-19  9:14 [PATCH] add http.maxReceiveSpeed to limit git-receive-pack receiving speed Xia XiaoWen
2021-08-19  9:36 ` Bagas Sanjaya
2021-08-19 11:27   ` Xiaowen Xia [this message]
2021-08-19 18:28 ` Junio C Hamano
2021-08-20 18:11   ` Jeff King
2021-08-20 18:21     ` Randall S. Becker
2021-08-21  3:19   ` Xiaowen Xia
2021-08-21  4:21     ` Junio C Hamano
2021-08-21  5:16       ` Xiaowen Xia
2021-08-23 16:04 ` [PATCH v2] http: add http.maxReceiveSpeed to limit receiving speed of "git-receive-pack" chenan.xxw
2021-08-24  0:31   ` Jiang Xin
2021-08-31  4:15     ` Xiaowen Xia

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='CAHLXgnZQmLiLbGnbC-nouE6tTJam5ZmdRmj=8E+tmM1pNLVFXw@mail.gmail.com' \
    --to=haoyurenzhuxia@gmail.com \
    --cc=bagasdotme@gmail.com \
    --cc=chenan.xxw@alibaba-inc.com \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=peff@peff.net \
    --cc=worldhello.net@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.