linux-kernel - Re: [patch v2 8/8]block: fiops sync preempts async

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAJd=RBC7YCFN3T98en5dBo9T0Jar9cU+=dFtTK96VQuqQtGfgw@mail.gmail.com>
Date:	Mon, 30 Jan 2012 19:53:41 +0800
From:	Hillf Danton <dhillf@...il.com>
To:	Shaohua Li <shaohua.li@...el.com>
Cc:	axboe@...nel.dk, linux-kernel@...r.kernel.org, vgoyal@...hat.com,
	david@...morbit.com, jack@...e.cz, zhu.yanhai@...il.com,
	namhyung.kim@....com
Subject: Re: [patch v2 8/8]block: fiops sync preempts async

Hello Shaohua

On Mon, Jan 30, 2012 at 3:02 PM, Shaohua Li <shaohua.li@...el.com> wrote:
> Debug patch only.
>
> This implements CFQ like sync preempts async. But like CFQ, this
> will starve async.
>
> ---
>  block/fiops-iosched.c |   15 +++++++++++++++
>  1 file changed, 15 insertions(+)
>
> Index: linux/block/fiops-iosched.c
> ===================================================================
> --- linux.orig/block/fiops-iosched.c    2012-01-21 10:35:04.000000000 +0800
> +++ linux/block/fiops-iosched.c 2012-01-21 10:48:52.000000000 +0800
> @@ -408,6 +408,21 @@ static struct fiops_ioc *fiops_select_io
>                return NULL;
>        }
>
> +       /* Let sync request preempt async queue */
> +       if (!rq_is_sync(rq) && service_tree->count > 1) {
> +               struct rb_node *tmp = rb_next(&ioc->rb_node);
> +               struct fiops_ioc *sync_ioc = NULL;
> +               while (tmp) {
> +                       sync_ioc = rb_entry(tmp, struct fiops_ioc, rb_node);
> +                       rq = rq_entry_fifo(sync_ioc->fifo.next);
> +                       if (rq_is_sync(rq))
> +                               break;
> +                       tmp = rb_next(&sync_ioc->rb_node);

			sync_ioc = NULL;        yes?
Hillf
> +               }
> +               if (sync_ioc)
> +                       ioc = sync_ioc;
> +       }
> +
>        return ioc;
>  }
>