mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Arnav Kapoor <kapoorarnav43@gmail.com>
To: Jonathan Cameron <jic23@kernel.org>
Cc: "David Lechner" <dlechner@baylibre.com>,
	"Nuno Sá" <nuno.sa@analog.com>,
	"Andy Shevchenko" <andy@kernel.org>,
	"Andy Shevchenko" <andriy.shevchenko@intel.com>,
	"Gwendal Grignou" <gwendal@chromium.org>,
	"Stephen Boyd" <swboyd@chromium.org>,
	linux-iio@vger.kernel.org, linux-kernel@vger.kernel.org,
	"Arnav Kapoor" <kapoorarnav43@gmail.com>
Subject: [PATCH 1/6] iio: proximity: sx_common: fix raw reads while the trigger is in use
Date: Tue, 17 Nov 2026 02:23:09 +0530	[thread overview]
Message-ID: <20261116205320.26070-2-kapoorarnav43@gmail.com> (raw)
In-Reply-To: <20261116205320.26070-1-kapoorarnav43@gmail.com>

sx_common_read_proximity() only reinitialises the completion after
waiting on it, so a complete() that arrives when no read is waiting is
consumed by the next read, which then returns old data without waiting
for a conversion that includes the channel it just enabled. This
happens in two ways:

- The device trigger has no validate_device(), so another IIO device
  can use it. While it does, set_trigger_state() keeps the CONVDONE
  interrupt enabled and the IRQ thread calls complete() for every
  conversion.
- A read interrupted by a signal reinitialises the completion and
  returns, but a CONVDONE interrupt already raised can still complete
  it once the mutex is dropped.

Also, when it is done, the read disables the CONVDONE interrupt
unconditionally, which stops the trigger for the other device.

Reinitialise the completion before enabling the interrupt, and leave
the interrupt enabled when the trigger is, as set_trigger_state()
already does for pending raw reads.

The interrupted-wait case was reported by the Sashiko AI reviewer on an
unrelated sx9324 patch.

Fixes: 72ad02b15d63 ("iio: Add SEMTECH SX9310/9311 sensor driver")
Signed-off-by: Arnav Kapoor <kapoorarnav43@gmail.com>
---
 drivers/iio/proximity/sx_common.c | 25 +++++++++++++++++++------
 1 file changed, 19 insertions(+), 6 deletions(-)

diff --git a/drivers/iio/proximity/sx_common.c b/drivers/iio/proximity/sx_common.c
index 4d4937395d9a..66d06435da05 100644
--- a/drivers/iio/proximity/sx_common.c
+++ b/drivers/iio/proximity/sx_common.c
@@ -167,6 +167,15 @@ static int sx_common_put_event_channel(struct sx_common_data *data, int channel)
 				     data->chan_event & ~BIT(channel));
 }
 
+/* The trigger needs the CONVDONE interrupt too, leave it on while in use. */
+static int sx_common_put_convdone_irq(struct sx_common_data *data)
+{
+	if (data->trigger_enabled)
+		return 0;
+
+	return sx_common_disable_irq(data, SX_COMMON_CONVDONE_IRQ);
+}
+
 /**
  * sx_common_read_proximity() - Read raw proximity value.
  * @data:	Internal data
@@ -188,18 +197,22 @@ int sx_common_read_proximity(struct sx_common_data *data,
 	if (ret)
 		goto out;
 
+	/*
+	 * Forget conversions that completed before this read, e.g. while the
+	 * trigger was in use, so that we wait for a fresh one.
+	 */
+	reinit_completion(&data->completion);
+
 	ret = sx_common_enable_irq(data, SX_COMMON_CONVDONE_IRQ);
 	if (ret)
 		goto out_put_channel;
 
 	mutex_unlock(&data->mutex);
 
-	if (data->client->irq) {
+	if (data->client->irq)
 		ret = wait_for_completion_interruptible(&data->completion);
-		reinit_completion(&data->completion);
-	} else {
+	else
 		ret = data->chip_info->ops.wait_for_sample(data);
-	}
 
 	mutex_lock(&data->mutex);
 
@@ -212,7 +225,7 @@ int sx_common_read_proximity(struct sx_common_data *data,
 
 	*val = sign_extend32(be16_to_cpu(rawval), chan->scan_type.realbits - 1);
 
-	ret = sx_common_disable_irq(data, SX_COMMON_CONVDONE_IRQ);
+	ret = sx_common_put_convdone_irq(data);
 	if (ret)
 		goto out_put_channel;
 
@@ -225,7 +238,7 @@ int sx_common_read_proximity(struct sx_common_data *data,
 	return IIO_VAL_INT;
 
 out_disable_irq:
-	sx_common_disable_irq(data, SX_COMMON_CONVDONE_IRQ);
+	sx_common_put_convdone_irq(data);
 out_put_channel:
 	sx_common_put_read_channel(data, chan->channel);
 out:
-- 
2.53.0


  reply	other threads:[~2026-11-16 20:53 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-16 20:53 [PATCH 0/6] iio: proximity: sx93xx: fix raw reads, thresholds, hysteresis and gain Arnav Kapoor
2026-11-16 20:53 ` Arnav Kapoor [this message]
2026-11-16 20:53 ` [PATCH 2/6] iio: proximity: sx9324: fix proximity threshold writes Arnav Kapoor
2026-11-16 20:53 ` [PATCH 3/6] iio: proximity: sx9324: reject negative hysteresis values Arnav Kapoor
2026-11-16 20:53 ` [PATCH 4/6] iio: proximity: sx9360: fix hardware gain encoding Arnav Kapoor
2026-11-16 20:53 ` [PATCH 5/6] iio: proximity: sx9360: fix proximity threshold writes Arnav Kapoor
2026-11-16 20:53 ` [PATCH 6/6] iio: proximity: sx9360: reject negative hysteresis values Arnav Kapoor

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=20261116205320.26070-2-kapoorarnav43@gmail.com \
    --to=kapoorarnav43@gmail.com \
    --cc=andriy.shevchenko@intel.com \
    --cc=andy@kernel.org \
    --cc=dlechner@baylibre.com \
    --cc=gwendal@chromium.org \
    --cc=jic23@kernel.org \
    --cc=linux-iio@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=nuno.sa@analog.com \
    --cc=swboyd@chromium.org \
    /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®