Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 05 Jan 2015 12:37:08 -0500
From:      John Baldwin <jhb@FreeBSD.org>
To:        Hans Petter Selasky <hselasky@FreeBSD.org>,  src-committers@freebsd.org, svn-src-all@freebsd.org,  svn-src-head@freebsd.org
Subject:   Re: svn commit: r276534 - head/sys/dev/vt
Message-ID:  <54AACBC4.5040802@FreeBSD.org>
In-Reply-To: <201501021335.t02DZBRm015072@svn.freebsd.org>
References:  <201501021335.t02DZBRm015072@svn.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 1/2/15 8:35 AM, Hans Petter Selasky wrote:
> Author: hselasky
> Date: Fri Jan  2 13:35:10 2015
> New Revision: 276534
> URL: https://svnweb.freebsd.org/changeset/base/276534
> 
> Log:
>   The "vt_suspend_flush_timer()" function is sometimes called locked
>   which prevents us from doing a "callout_drain()" call. The callout in
>   question has a lock associated with it and we are not freeing the
>   callout. That means we can use the "callout_stop()" function to
>   atomically stop the callback iff the "callout_stop()" function is
>   called locked. This patch applies proper locking to "callout_stop()"
>   and replaces a "callout_drain()" with a "callout_stop()".
>
> Modified: head/sys/dev/vt/vt_core.c
> ==============================================================================
> --- head/sys/dev/vt/vt_core.c	Fri Jan  2 13:15:36 2015	(r276533)
> +++ head/sys/dev/vt/vt_core.c	Fri Jan  2 13:35:10 2015	(r276534)
> @@ -114,6 +114,7 @@ const struct terminal_class vt_termclass
>  
>  #define	VT_LOCK(vd)	mtx_lock(&(vd)->vd_lock)
>  #define	VT_UNLOCK(vd)	mtx_unlock(&(vd)->vd_lock)
> +#define	VT_LOCK_ASSERT(vd, what)	mtx_assert(&(vd)->vd_lock, what)
>  
>  #define	VT_UNIT(vw)	((vw)->vw_device->vd_unit * VT_MAXWINDOWS + \
>  			(vw)->vw_number)
> @@ -283,12 +284,18 @@ vt_resume_flush_timer(struct vt_device *
>  static void
>  vt_suspend_flush_timer(struct vt_device *vd)
>  {
> +	/*
> +	 * As long as this function is called locked, callout_stop()
> +	 * has the same effect like callout_drain() with regard to
> +	 * preventing the callback function from executing.
> +	 */
> +	VT_LOCK_ASSERT(vd, MA_OWNED);

Note that this assert is redundant.  callout_stop() already asserts this
for a callout initialized via callout_init_mtx().

>  
>  	if (!(vd->vd_flags & VDF_ASYNC) ||
>  	    !atomic_cmpset_int(&vd->vd_timer_armed, 1, 0))
>  		return;
>  
> -	callout_drain(&vd->vd_timer);
> +	callout_stop(&vd->vd_timer);
>  }
>  
>  static void


-- 
John Baldwin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?54AACBC4.5040802>