mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Matthew Laux <matthew.laux@gmail.com>
To: Mauro Carvalho Chehab <mchehab@kernel.org>,
	Hans Verkuil <hverkuil@kernel.org>,
	Mehdi Djait <mehdi.djait@bootlin.com>
Cc: linux-kernel@vger.kernel.org, linux-media@vger.kernel.org,
	Matthew Laux <matthew.laux@gmail.com>
Subject: [PATCH v7 08/12] media: tw9900: use runtime PM autosuspend
Date: Fri, 20 Nov 2026 21:55:19 -0600	[thread overview]
Message-ID: <20261121035523.26764-9-matthew.laux@gmail.com> (raw)
In-Reply-To: <20261121035523.26764-1-matthew.laux@gmail.com>

Each runtime resume enables the supply, toggles reset, rewrites the
init table and waits 300 ms for the horizontal sync PLL to lock. Without
autosuspend, a typical chain of s_routing, polling g_input_status,
querystd, and s_stream would suspend and repower the device each
operation, which adds ~350 ms per call and discards the signal lock.

Use autosuspend with a 1sec delay so that closely spaced operations
share a single power-up.

Signed-off-by: Matthew Laux <matthew.laux@gmail.com>
---
 drivers/media/i2c/tw9900.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/media/i2c/tw9900.c b/drivers/media/i2c/tw9900.c
index b2b754e898c2..b2c30e84acb8 100644
--- a/drivers/media/i2c/tw9900.c
+++ b/drivers/media/i2c/tw9900.c
@@ -68,6 +68,7 @@
 #define TW9900_STD_AUTO				7
 
 #define TW9900_VIDEO_POLL_TRIES			20
+#define TW9900_AUTOSUSPEND_DELAY_MS		1000
 
 struct regval {
 	u8 addr;
@@ -346,7 +347,7 @@ static int tw9900_s_stream(struct v4l2_subdev *sd, int on)
 
 		mutex_unlock(&tw9900->mutex);
 
-		pm_runtime_put(&client->dev);
+		pm_runtime_put_autosuspend(&client->dev);
 
 		return ret;
 	}
@@ -355,7 +356,7 @@ static int tw9900_s_stream(struct v4l2_subdev *sd, int on)
 
 err_unlock:
 	mutex_unlock(&tw9900->mutex);
-	pm_runtime_put(&client->dev);
+	pm_runtime_put_autosuspend(&client->dev);
 
 	return ret;
 }
@@ -532,7 +533,7 @@ static int tw9900_querystd(struct v4l2_subdev *sd, v4l2_std_id *std)
 
 out_unlock:
 	mutex_unlock(&tw9900->mutex);
-	pm_runtime_put(&tw9900->client->dev);
+	pm_runtime_put_autosuspend(&tw9900->client->dev);
 
 	return ret;
 }
@@ -568,7 +569,7 @@ static int tw9900_g_input_status(struct v4l2_subdev *sd, u32 *status)
 	ret = tw9900_read_reg(tw9900->client, TW9900_REG_CHIP_STATUS);
 	mutex_unlock(&tw9900->mutex);
 
-	pm_runtime_put(&tw9900->client->dev);
+	pm_runtime_put_autosuspend(&tw9900->client->dev);
 
 	if (ret < 0)
 		return ret;
@@ -623,7 +624,7 @@ static int tw9900_check_id(struct tw9900 *tw9900,
 	ret = tw9900_read_reg(client, TW9900_REG_CHIP_ID);
 	mutex_unlock(&tw9900->mutex);
 
-	pm_runtime_put(&tw9900->client->dev);
+	pm_runtime_put_autosuspend(&tw9900->client->dev);
 
 	if (ret < 0)
 		return ret;
@@ -712,6 +713,7 @@ static int tw9900_runtime_suspend(struct device *dev)
 
 static void tw9900_pm_cleanup(struct device *dev)
 {
+	pm_runtime_dont_use_autosuspend(dev);
 	pm_runtime_disable(dev);
 	if (!pm_runtime_status_suspended(dev))
 		tw9900_runtime_suspend(dev);
@@ -782,6 +784,8 @@ static int tw9900_probe(struct i2c_client *client)
 		goto err_free_handler;
 
 	pm_runtime_set_suspended(dev);
+	pm_runtime_set_autosuspend_delay(dev, TW9900_AUTOSUSPEND_DELAY_MS);
+	pm_runtime_use_autosuspend(dev);
 	pm_runtime_enable(dev);
 
 	ret = tw9900_check_id(tw9900, client);
-- 
2.47.3


  parent reply	other threads:[~2026-11-21  3:55 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-21  3:55 [PATCH v7 00/12] Add driver for StarTech USB3HDCAP capture device Matthew Laux
2026-11-21  3:55 ` [PATCH v7 01/12] media: tw9900: Disable regulator on runtime resume failure Matthew Laux
2026-11-21  3:55 ` [PATCH v7 02/12] media: tw9900: release the device when stopping the stream fails Matthew Laux
2026-11-21  3:55 ` [PATCH v7 03/12] media: tw9900: accept the TW9910 chip ID Matthew Laux
2026-11-21  3:55 ` [PATCH v7 04/12] media: tw9900: add saturation, hue and sharpness controls Matthew Laux
2026-11-21  3:55 ` [PATCH v7 05/12] media: tw9900: program all mode registers in the PAL table Matthew Laux
2026-11-21  3:55 ` [PATCH v7 06/12] media: tw9900: power off the device on remove if still active Matthew Laux
2026-11-21  3:55 ` [PATCH v7 07/12] media: tw9900: power down analog blocks in runtime suspend Matthew Laux
2026-11-21  3:55 ` Matthew Laux [this message]
2026-11-21  3:55 ` [PATCH v7 09/12] media: tw9900: allow board-specific register tables via platform data Matthew Laux
2026-11-21  3:55 ` [PATCH v7 10/12] media: tw9900: add S-Video input and progressive source detection Matthew Laux
2026-11-21  3:55 ` [PATCH v7 11/12] media: Add subdevice driver for the Mstar MST3367 HDMI receiver Matthew Laux
2026-11-21  3:55 ` [PATCH v7 12/12] media: add driver for StarTech USB3HDCAP Matthew Laux

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=20261121035523.26764-9-matthew.laux@gmail.com \
    --to=matthew.laux@gmail.com \
    --cc=hverkuil@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=mehdi.djait@bootlin.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

all inboxes | Powered by JetHome®