All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: Xu Qiang <xuqiang36@huawei.com>
Cc: tzimmermann@suse.de, airlied@linux.ie,
	linux-kernel@vger.kernel.org, rui.xiang@huawei.com,
	dri-devel@lists.freedesktop.org
Subject: Re: [PATCH -next] gpu: drm: Fix spinlock vblank_time_lock use error.
Date: Wed, 22 Jul 2020 14:40:56 +0300	[thread overview]
Message-ID: <20200722114056.GB6112@intel.com> (raw)
In-Reply-To: <20200722010527.52293-1-xuqiang36@huawei.com>

On Wed, Jul 22, 2020 at 01:05:27AM +0000, Xu Qiang wrote:
> The drm_handle_vblank function is in the interrupt context.
> Therefore, the spin lock vblank_time_lock is obtained
> from the interrupt context.
> 
> Cc: <stable@vger.kernel.org>
> Signed-off-by: Xu Qiang <xuqiang36@huawei.com>
> ---
>  drivers/gpu/drm/drm_vblank.c | 17 ++++++++++-------
>  1 file changed, 10 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
> index f402c75b9d34..4ca63ff33a43 100644
> --- a/drivers/gpu/drm/drm_vblank.c
> +++ b/drivers/gpu/drm/drm_vblank.c
> @@ -229,10 +229,11 @@ static void drm_reset_vblank_timestamp(struct drm_device *dev, unsigned int pipe
>  {
>  	u32 cur_vblank;
>  	bool rc;
> +	unsigned long irqflags;
>  	ktime_t t_vblank;
>  	int count = DRM_TIMESTAMP_MAXRETRIES;
>  
> -	spin_lock(&dev->vblank_time_lock);
> +	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);

Nak. This is always called with interrupts off, so no point on wasting
time saving/restoring the flags. And it's the same situation for all the
other cases you have below.

>  
>  	/*
>  	 * sample the current counter to avoid random jumps
> @@ -257,7 +258,7 @@ static void drm_reset_vblank_timestamp(struct drm_device *dev, unsigned int pipe
>  	 */
>  	store_vblank(dev, pipe, 1, t_vblank, cur_vblank);
>  
> -	spin_unlock(&dev->vblank_time_lock);
> +	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
>  }
>  
>  /*
> @@ -1106,11 +1107,12 @@ static int __enable_vblank(struct drm_device *dev, unsigned int pipe)
>  static int drm_vblank_enable(struct drm_device *dev, unsigned int pipe)
>  {
>  	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
> +	unsigned long irqflags;
>  	int ret = 0;
>  
>  	assert_spin_locked(&dev->vbl_lock);
>  
> -	spin_lock(&dev->vblank_time_lock);
> +	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);
>  
>  	if (!vblank->enabled) {
>  		/*
> @@ -1136,7 +1138,7 @@ static int drm_vblank_enable(struct drm_device *dev, unsigned int pipe)
>  		}
>  	}
>  
> -	spin_unlock(&dev->vblank_time_lock);
> +	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
>  
>  	return ret;
>  }
> @@ -1917,6 +1919,7 @@ bool drm_handle_vblank(struct drm_device *dev, unsigned int pipe)
>  {
>  	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
>  	unsigned long irqflags;
> +	unsigned long irqflags_vblank;
>  	bool disable_irq;
>  
>  	if (drm_WARN_ON_ONCE(dev, !drm_dev_has_vblank(dev)))
> @@ -1931,18 +1934,18 @@ bool drm_handle_vblank(struct drm_device *dev, unsigned int pipe)
>  	 * vblank enable/disable, as this would cause inconsistent
>  	 * or corrupted timestamps and vblank counts.
>  	 */
> -	spin_lock(&dev->vblank_time_lock);
> +	spin_lock_irqsave(&dev->vblank_time_lock, irqflags_vblank);
>  
>  	/* Vblank irq handling disabled. Nothing to do. */
>  	if (!vblank->enabled) {
> -		spin_unlock(&dev->vblank_time_lock);
> +		spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags_vblank);
>  		spin_unlock_irqrestore(&dev->event_lock, irqflags);
>  		return false;
>  	}
>  
>  	drm_update_vblank_count(dev, pipe, true);
>  
> -	spin_unlock(&dev->vblank_time_lock);
> +	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags_vblank);
>  
>  	wake_up(&vblank->queue);
>  
> -- 
> 2.25.0
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Ville Syrjälä
Intel
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

WARNING: multiple messages have this Message-ID (diff)
From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: Xu Qiang <xuqiang36@huawei.com>
Cc: maarten.lankhorst@linux.intel.com, mripard@kernel.org,
	tzimmermann@suse.de, airlied@linux.ie, daniel@ffwll.ch,
	rui.xiang@huawei.com, linux-kernel@vger.kernel.org,
	dri-devel@lists.freedesktop.org
Subject: Re: [PATCH -next] gpu: drm: Fix spinlock vblank_time_lock use error.
Date: Wed, 22 Jul 2020 14:40:56 +0300	[thread overview]
Message-ID: <20200722114056.GB6112@intel.com> (raw)
In-Reply-To: <20200722010527.52293-1-xuqiang36@huawei.com>

On Wed, Jul 22, 2020 at 01:05:27AM +0000, Xu Qiang wrote:
> The drm_handle_vblank function is in the interrupt context.
> Therefore, the spin lock vblank_time_lock is obtained
> from the interrupt context.
> 
> Cc: <stable@vger.kernel.org>
> Signed-off-by: Xu Qiang <xuqiang36@huawei.com>
> ---
>  drivers/gpu/drm/drm_vblank.c | 17 ++++++++++-------
>  1 file changed, 10 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
> index f402c75b9d34..4ca63ff33a43 100644
> --- a/drivers/gpu/drm/drm_vblank.c
> +++ b/drivers/gpu/drm/drm_vblank.c
> @@ -229,10 +229,11 @@ static void drm_reset_vblank_timestamp(struct drm_device *dev, unsigned int pipe
>  {
>  	u32 cur_vblank;
>  	bool rc;
> +	unsigned long irqflags;
>  	ktime_t t_vblank;
>  	int count = DRM_TIMESTAMP_MAXRETRIES;
>  
> -	spin_lock(&dev->vblank_time_lock);
> +	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);

Nak. This is always called with interrupts off, so no point on wasting
time saving/restoring the flags. And it's the same situation for all the
other cases you have below.

>  
>  	/*
>  	 * sample the current counter to avoid random jumps
> @@ -257,7 +258,7 @@ static void drm_reset_vblank_timestamp(struct drm_device *dev, unsigned int pipe
>  	 */
>  	store_vblank(dev, pipe, 1, t_vblank, cur_vblank);
>  
> -	spin_unlock(&dev->vblank_time_lock);
> +	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
>  }
>  
>  /*
> @@ -1106,11 +1107,12 @@ static int __enable_vblank(struct drm_device *dev, unsigned int pipe)
>  static int drm_vblank_enable(struct drm_device *dev, unsigned int pipe)
>  {
>  	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
> +	unsigned long irqflags;
>  	int ret = 0;
>  
>  	assert_spin_locked(&dev->vbl_lock);
>  
> -	spin_lock(&dev->vblank_time_lock);
> +	spin_lock_irqsave(&dev->vblank_time_lock, irqflags);
>  
>  	if (!vblank->enabled) {
>  		/*
> @@ -1136,7 +1138,7 @@ static int drm_vblank_enable(struct drm_device *dev, unsigned int pipe)
>  		}
>  	}
>  
> -	spin_unlock(&dev->vblank_time_lock);
> +	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags);
>  
>  	return ret;
>  }
> @@ -1917,6 +1919,7 @@ bool drm_handle_vblank(struct drm_device *dev, unsigned int pipe)
>  {
>  	struct drm_vblank_crtc *vblank = &dev->vblank[pipe];
>  	unsigned long irqflags;
> +	unsigned long irqflags_vblank;
>  	bool disable_irq;
>  
>  	if (drm_WARN_ON_ONCE(dev, !drm_dev_has_vblank(dev)))
> @@ -1931,18 +1934,18 @@ bool drm_handle_vblank(struct drm_device *dev, unsigned int pipe)
>  	 * vblank enable/disable, as this would cause inconsistent
>  	 * or corrupted timestamps and vblank counts.
>  	 */
> -	spin_lock(&dev->vblank_time_lock);
> +	spin_lock_irqsave(&dev->vblank_time_lock, irqflags_vblank);
>  
>  	/* Vblank irq handling disabled. Nothing to do. */
>  	if (!vblank->enabled) {
> -		spin_unlock(&dev->vblank_time_lock);
> +		spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags_vblank);
>  		spin_unlock_irqrestore(&dev->event_lock, irqflags);
>  		return false;
>  	}
>  
>  	drm_update_vblank_count(dev, pipe, true);
>  
> -	spin_unlock(&dev->vblank_time_lock);
> +	spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags_vblank);
>  
>  	wake_up(&vblank->queue);
>  
> -- 
> 2.25.0
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Ville Syrjälä
Intel

  reply	other threads:[~2020-07-22 11:41 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-22  1:05 [PATCH -next] gpu: drm: Fix spinlock vblank_time_lock use error Xu Qiang
2020-07-22  1:05 ` Xu Qiang
2020-07-22 11:40 ` Ville Syrjälä [this message]
2020-07-22 11:40   ` Ville Syrjälä

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=20200722114056.GB6112@intel.com \
    --to=ville.syrjala@linux.intel.com \
    --cc=airlied@linux.ie \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rui.xiang@huawei.com \
    --cc=tzimmermann@suse.de \
    --cc=xuqiang36@huawei.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.