Linux-ARM-Kernel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Mehdi Djait <mehdi.djait@linux.intel.com>
To: Daniel Golle <daniel@makrotopia.org>
Cc: Michael Riesch <michael.riesch@collabora.com>,
	 Mauro Carvalho Chehab <mchehab@kernel.org>,
	Heiko Stuebner <heiko@sntech.de>,
	 Hans Verkuil <hverkuil+cisco@kernel.org>,
	Bryan O'Donoghue <bryan.odonoghue@linaro.org>,
	 Sakari Ailus <sakari.ailus@linux.intel.com>,
	linux-media@vger.kernel.org,
	 linux-arm-kernel@lists.infradead.org,
	linux-rockchip@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/6] media: rockchip: rkcif: fix the MIPI stop path locking
Date: Mon, 16 Nov 2026 17:25:52 +0100	[thread overview]
Message-ID: <avstwwwxx0KrGe23@mdjait-mobl> (raw)
In-Reply-To: <744eeec346f96d580a9215c657951f5a737e734e.1793702189.git.daniel@makrotopia.org>

Hello Daniel,

Thank you for the patch!

On Tue, Nov 03, 2026 at 10:45:29AM +0000, Daniel Golle wrote:
> rkcif_mipi_stop_streaming() takes the subdev state mutex, and the
> interrupt handler calls it at the frame end that completes a stop. A
> mutex taken in hard interrupt context sleeps when it is contended, and
> CONFIG_DEBUG_ATOMIC_SLEEP reports the call on every stream stop. The
> state itself is not read on either path; the lock serialises the
> read-modify-write of the interrupt enable and status registers, which
> the streams of a MIPI block share.
> 
> Guard those registers with a spinlock per interface, taken with
> interrupts disabled, so that the handler and the process context paths
> exclude each other, and leave the subdev state alone.
> 
> Fixes: 1f2353f5a1af ("media: rockchip: rkcif: add support for rk3568 vicap mipi capture")
> Assisted-by: LLM
> Signed-off-by: Daniel Golle <daniel@makrotopia.org>
> ---
>  .../rockchip/rkcif/rkcif-capture-mipi.c       | 24 +++++++------------
>  .../platform/rockchip/rkcif/rkcif-common.h    |  3 +++
>  2 files changed, 12 insertions(+), 15 deletions(-)
> 
> diff --git a/drivers/media/platform/rockchip/rkcif/rkcif-capture-mipi.c b/drivers/media/platform/rockchip/rkcif/rkcif-capture-mipi.c
> index 686b995525b9b..53cfd37bf02bd 100644
> --- a/drivers/media/platform/rockchip/rkcif/rkcif-capture-mipi.c
> +++ b/drivers/media/platform/rockchip/rkcif/rkcif-capture-mipi.c
> @@ -721,17 +721,14 @@ static int rkcif_mipi_start_streaming(struct rkcif_stream *stream)
>  	struct rkcif_interface *interface = stream->interface;
>  	const struct rkcif_output_fmt *active_out_fmt;
>  	const struct rkcif_mipi_match_data *match_data;
> -	struct v4l2_subdev_state *state;
>  	u32 ctrl0 = 0, ctrl1 = 0, int_temp = 0, int_mask = 0, vlw = 0;
> +	unsigned long flags;
>  	u16 height, width;
> -	int ret = -EINVAL;
> -
> -	state = v4l2_subdev_lock_and_get_active_state(&interface->sd);
>  
>  	active_out_fmt = rkcif_stream_find_output_fmt(stream, false,
>  						      stream->pix.pixelformat);
>  	if (!active_out_fmt)
> -		goto out;
> +		return -EINVAL;
>  
>  	height = stream->pix.height;
>  	width = stream->pix.width;
> @@ -746,6 +743,7 @@ static int rkcif_mipi_start_streaming(struct rkcif_stream *stream)
>  	int_mask |= RKCIF_MIPI_INT_FRAME0_END(stream->id);
>  	int_mask |= RKCIF_MIPI_INT_FRAME1_END(stream->id);
>  
> +	spin_lock_irqsave(&interface->irq_lock, flags);
>  	int_temp = rkcif_mipi_read(interface, RKCIF_MIPI_INTEN);
>  	int_temp |= int_mask;
>  	rkcif_mipi_write(interface, RKCIF_MIPI_INTEN, int_temp);
> @@ -753,6 +751,7 @@ static int rkcif_mipi_start_streaming(struct rkcif_stream *stream)
>  	int_temp = rkcif_mipi_read(interface, RKCIF_MIPI_INTSTAT);
>  	int_temp &= ~int_mask;
>  	rkcif_mipi_write(interface, RKCIF_MIPI_INTSTAT, int_temp);
> +	spin_unlock_irqrestore(&interface->irq_lock, flags);
>  
>  	rkcif_mipi_stream_write(stream, RKCIF_MIPI_FRAME0_VLW_Y, vlw);
>  	rkcif_mipi_stream_write(stream, RKCIF_MIPI_FRAME1_VLW_Y, vlw);
> @@ -769,26 +768,21 @@ static int rkcif_mipi_start_streaming(struct rkcif_stream *stream)
>  	 */
>  	rkcif_mipi_write(interface, RKCIF_MIPI_CTRL, RK3588_MIPI_CTRL_CAP_EN);
>  
> -	ret = 0;
> -
> -out:
> -	v4l2_subdev_unlock_state(state);
> -	return ret;
> +	return 0;
>  }
>  
>  static void rkcif_mipi_stop_streaming(struct rkcif_stream *stream)
>  {
>  	struct rkcif_interface *interface = stream->interface;
> -	struct v4l2_subdev_state *state;
>  	u32 int_temp = 0, int_mask = 0;
> -
> -	state = v4l2_subdev_lock_and_get_active_state(&interface->sd);
> +	unsigned long flags;
>  
>  	rkcif_mipi_stream_write(stream, RKCIF_MIPI_CTRL0, 0);
>  
>  	int_mask |= RKCIF_MIPI_INT_FRAME0_END(stream->id);
>  	int_mask |= RKCIF_MIPI_INT_FRAME1_END(stream->id);
>  
> +	spin_lock_irqsave(&interface->irq_lock, flags);
>  	int_temp = rkcif_mipi_read(interface, RKCIF_MIPI_INTEN);
>  	int_temp &= ~int_mask;
>  	rkcif_mipi_write(interface, RKCIF_MIPI_INTEN, int_temp);
> @@ -796,10 +790,9 @@ static void rkcif_mipi_stop_streaming(struct rkcif_stream *stream)
>  	int_temp = rkcif_mipi_read(interface, RKCIF_MIPI_INTSTAT);
>  	int_temp &= ~int_mask;
>  	rkcif_mipi_write(interface, RKCIF_MIPI_INTSTAT, int_temp);
> +	spin_unlock_irqrestore(&interface->irq_lock, flags);
>  
>  	stream->stopping = false;
> -
> -	v4l2_subdev_unlock_state(state);
>  }
>  
>  static void rkcif_mipi_set_crop(struct rkcif_stream *stream, u16 left, u16 top)
> @@ -881,6 +874,7 @@ int rkcif_mipi_register(struct rkcif_device *rkcif)
>  		interface->in_fmts_num = ARRAY_SIZE(mipi_in_fmts);
>  		interface->set_crop = rkcif_mipi_set_crop;
>  		interface->streams_num = 0;
> +		spin_lock_init(&interface->irq_lock);
>  		ret = rkcif_interface_register(rkcif, interface);
>  		if (ret)
>  			continue;
> diff --git a/drivers/media/platform/rockchip/rkcif/rkcif-common.h b/drivers/media/platform/rockchip/rkcif/rkcif-common.h
> index 4d9211ba9bda8..16a5ecb25c2fb 100644
> --- a/drivers/media/platform/rockchip/rkcif/rkcif-common.h
> +++ b/drivers/media/platform/rockchip/rkcif/rkcif-common.h
> @@ -198,6 +198,9 @@ struct rkcif_interface {
>  		struct rkcif_dvp dvp;
>  	};
>  
> +	/* interrupt enable and status are shared by the streams of a block */
> +	spinlock_t irq_lock;
> +
>  	void (*set_crop)(struct rkcif_stream *stream, u16 left, u16 top);
>  };

Could you please add this to this patch to fix the timeout issue when
stopping the stream: I sent a patch series to fix some issues on the
rkcif driver.

I fixed also the stop stream timeout issue and then I saw
that it is quite similar to your patch.

You can give this a test and tell me how it goes:

$ git diff
diff --git a/drivers/media/platform/rockchip/rkcif/rkcif-stream.c b/drivers/media/platform/rockchip/rkcif/rkcif-stream.c
index 66726f654674..c72e22e0d726 100644
--- a/drivers/media/platform/rockchip/rkcif/rkcif-stream.c
+++ b/drivers/media/platform/rockchip/rkcif/rkcif-stream.c
@@ -308,10 +308,6 @@ static void rkcif_stream_stop_streaming(struct vb2_queue *queue)
        u64 mask;
        int ret;
 
-       mask = BIT_ULL(stream->id);
-       v4l2_subdev_disable_streams(&stream->interface->sd, RKCIF_IF_PAD_SRC,
-                                   mask);
-
        stream->stopping = true;
        ret = wait_event_timeout(stream->wq_stopped, !stream->stopping,
                                 msecs_to_jiffies(1000));
@@ -319,6 +315,10 @@ static void rkcif_stream_stop_streaming(struct vb2_queue *queue)
        if (!ret && stream->stop_streaming)
                stream->stop_streaming(stream);
 
+       mask = BIT_ULL(stream->id);
+       v4l2_subdev_disable_streams(&stream->interface->sd, RKCIF_IF_PAD_SRC,
+                                   mask);
+
        pm_runtime_put(rkcif->dev);
 
        rkcif_stream_return_all_buffers(stream, VB2_BUF_STATE_ERROR);

--
Kind Regards
Mehdi Djait




  reply	other threads:[~2026-11-16 16:26 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-03 10:45 [PATCH 0/6] media: rockchip: rkcif: recover from a lost MIPI line Daniel Golle
2026-11-03 10:45 ` [PATCH 1/6] media: rockchip: rkcif: mask MIPI interrupts when powered Daniel Golle
2026-11-03 13:04   ` Michael Riesch
2026-11-03 10:45 ` [PATCH 2/6] media: rockchip: rkcif: fix the MIPI stop path locking Daniel Golle
2026-11-16 16:25   ` Mehdi Djait [this message]
2026-11-03 10:45 ` [PATCH 3/6] media: rockchip: rkcif: count the MIPI capture errors Daniel Golle
2026-11-03 10:45 ` [PATCH 4/6] media: rockchip: rkcif: restart the MIPI DMA when it slips Daniel Golle
2026-11-03 10:45 ` [PATCH 5/6] media: rockchip: rkcif: raise the MIPI DMA write priority Daniel Golle
2026-11-03 10:45 ` [PATCH 6/6] media: synopsys: csi2rx: handle the error interrupts Daniel Golle

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=avstwwwxx0KrGe23@mdjait-mobl \
    --to=mehdi.djait@linux.intel.com \
    --cc=bryan.odonoghue@linaro.org \
    --cc=daniel@makrotopia.org \
    --cc=heiko@sntech.de \
    --cc=hverkuil+cisco@kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=linux-rockchip@lists.infradead.org \
    --cc=mchehab@kernel.org \
    --cc=michael.riesch@collabora.com \
    --cc=sakari.ailus@linux.intel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox