From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jeff Garzik <jeff@garzik.org>
Subject: Re: [Patch] CLD: fix hang in ncld_sess_open if session open fails
Date: Wed, 14 Apr 2010 00:01:37 -0400
Message-ID: <4BC53E21.2050600@garzik.org>
References: <20100413182214.28c6b565@redhat.com>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Return-path: <hail-devel-owner@vger.kernel.org>
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
        d=gmail.com; s=gamma;
        h=domainkey-signature:received:received:sender:message-id:date:from
         :user-agent:mime-version:to:cc:subject:references:in-reply-to
         :content-type:content-transfer-encoding;
        bh=fUSBS0DzLyB2CPj5JiH5AyhmhiK+TU+7iAa+7lfiJ5s=;
        b=HyUQbBfovpFtXQuaKWgL8U4l8amkJimWRPga5+WFApJXH/7neC0wFvOhvVTJpXqZSP
         C3k4G8mAwPEAbHOi/49fp5G/luWed8NHFXFK5hwyn5UTLbIL6fqzPHsbPbsXvtuQp+io
         zHEwcbqGJwGxz8BksZQXJa9h+wwc5butEOqOQ=
In-Reply-To: <20100413182214.28c6b565@redhat.com>
Sender: hail-devel-owner@vger.kernel.org
List-ID: <hail-devel.vger.kernel.org>
Content-Type: text/plain; charset="us-ascii"; format="flowed"
To: Pete Zaitcev <zaitcev@redhat.com>
Cc: Project Hail List <hail-devel@vger.kernel.org>

On 04/13/2010 08:22 PM, Pete Zaitcev wrote:
> The problem turned out to be two-fold, with the same symptom.
>
> Firstly, we use is_open to signal that the caller thread may
> proceed, but this is incorrect in case the thread open fails:
> we still want the caller thread to proceed and deliver the
> error indicator from ncld_sess_open to the application.
> So, let's split is_up from the condition variable mechanism.
> It continues to mean that the session is open and up, and
> open_done is introduced for the waiting mechanism.
>
> In addition, we forgot to take a mutex around a call into
> the cldc layer. It manifested itself in timers not firing,
> and so we would hang waiting for an answer from CLD server.
>
> Signed-off-by: Pete Zaitcev<zaitcev@redhat.com>
>
> ---
>   include/ncld.h |    1 +
>   lib/cldc.c     |   15 +++++++++++++--
>   2 files changed, 14 insertions(+), 2 deletions(-)
>
> This, I think, can go either before or after the e2e verbose.
> My previous patch added a HAIL_VERBOSE in a strategic location,
> but it's not really necessary and I took it out.

applied