From: David Miller <davem@davemloft.net>
To: billfink@mindspring.com
Cc: ilpo.jarvinen@helsinki.fi, netdev@vger.kernel.org,
herbert@gondor.apana.org.au, jheffner@psc.edu
Subject: Re: TSO trimming question
Date: Fri, 21 Dec 2007 01:27:20 -0800 (PST) [thread overview]
Message-ID: <20071221.012720.67812056.davem@davemloft.net> (raw)
In-Reply-To: <20071221030648.389669c4.billfink@mindspring.com>
From: Bill Fink <billfink@mindspring.com>
Date: Fri, 21 Dec 2007 03:06:48 -0500
> What's with all the shifting back and forth? Here with:
>
> ((jiffies<<1)>>1) - (tp->tso_deferred>>1)
>
> and later with:
>
> /* Ok, it looks like it is advisable to defer. */
> tp->tso_deferred = 1 | (jiffies<<1);
>
> Is this just done to try and avoid the special case of jiffies==0
> when the jiffies wrap? If so it seems like a lot of unnecessary
> work just to avoid a 1 in 4 billion event, since it's my understanding
> that the whole tcp_tso_should_defer function is just an optimization
> and not a criticality to the proper functioning of TCP, especially
> considering it hasn't even been executing at all up to now.
How else would you avoid the incorrect result when jiffies is
indeed zero?
It's two shifts, and this gets scheduled along with the other
instructions on many cpus so it's effectively free.
I don't see why this is even worth mentioning and discussing.
next prev parent reply other threads:[~2007-12-21 9:27 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-12-19 21:46 TSO trimming question Ilpo Järvinen
2007-12-19 22:02 ` Ilpo Järvinen
2007-12-20 3:26 ` Herbert Xu
2007-12-20 7:55 ` David Miller
2007-12-20 7:54 ` David Miller
2007-12-20 11:40 ` Ilpo Järvinen
2007-12-20 11:56 ` David Miller
2007-12-20 16:02 ` John Heffner
2007-12-21 4:36 ` David Miller
2007-12-21 8:06 ` Bill Fink
2007-12-21 9:26 ` Ilpo Järvinen
2007-12-21 9:28 ` Ilpo Järvinen
2007-12-21 9:27 ` David Miller [this message]
2007-12-21 9:29 ` Herbert Xu
2007-12-21 9:36 ` David Miller
2007-12-21 10:58 ` Bill Fink
2007-12-21 18:54 ` Bill Fink
2007-12-21 18:58 ` Ilpo Järvinen
2007-12-21 19:37 ` Bill Fink
2007-12-20 12:00 ` David Miller
2007-12-20 12:35 ` Ilpo Järvinen
2007-12-20 14:00 ` Herbert Xu
2007-12-20 23:55 ` David Miller
2007-12-21 18:55 ` [PATCH] [TCP]: Force TSO splits to MSS boundaries Ilpo Järvinen
2007-12-21 20:01 ` Ilpo Järvinen
2007-12-25 5:35 ` David Miller
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=20071221.012720.67812056.davem@davemloft.net \
--to=davem@davemloft.net \
--cc=billfink@mindspring.com \
--cc=herbert@gondor.apana.org.au \
--cc=ilpo.jarvinen@helsinki.fi \
--cc=jheffner@psc.edu \
--cc=netdev@vger.kernel.org \
/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).