linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bartosz Golaszewski <brgl@bgdev.pl>
To: Kent Gibson <warthog618@gmail.com>,
	Linus Walleij <linus.walleij@linaro.org>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: linux-gpio@vger.kernel.org, linux-kernel@vger.kernel.org,
	Bartosz Golaszewski <bgolaszewski@baylibre.com>
Subject: [RESEND PATCH v6 3/7] gpiolib: rework the locking mechanism for lineevent kfifo
Date: Tue, 11 Feb 2020 10:19:33 +0100	[thread overview]
Message-ID: <20200211091937.29558-4-brgl@bgdev.pl> (raw)
In-Reply-To: <20200211091937.29558-1-brgl@bgdev.pl>

From: Bartosz Golaszewski <bgolaszewski@baylibre.com>

The read_lock mutex is supposed to prevent collisions between reading
and writing to the line event kfifo but it's actually only taken when
the events are being read from it.

Drop the mutex entirely and reuse the spinlock made available to us in
the waitqueue struct. Take the lock whenever the fifo is modified or
inspected. Drop the call to kfifo_to_user() and instead first extract
the new element from kfifo when the lock is taken and only then pass
it on to the user after the spinlock is released.

Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 drivers/gpio/gpiolib.c | 64 +++++++++++++++++++++++-------------------
 1 file changed, 35 insertions(+), 29 deletions(-)

diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 753283486037..43d98309e725 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -787,8 +787,6 @@ static int linehandle_create(struct gpio_device *gdev, void __user *ip)
  * @irq: the interrupt that trigger in response to events on this GPIO
  * @wait: wait queue that handles blocking reads of events
  * @events: KFIFO for the GPIO events
- * @read_lock: mutex lock to protect reads from colliding with adding
- * new events to the FIFO
  * @timestamp: cache for the timestamp storing it between hardirq
  * and IRQ thread, used to bring the timestamp close to the actual
  * event
@@ -801,7 +799,6 @@ struct lineevent_state {
 	int irq;
 	wait_queue_head_t wait;
 	DECLARE_KFIFO(events, struct gpioevent_data, 16);
-	struct mutex read_lock;
 	u64 timestamp;
 };
 
@@ -817,7 +814,7 @@ static __poll_t lineevent_poll(struct file *filep,
 
 	poll_wait(filep, &le->wait, wait);
 
-	if (!kfifo_is_empty(&le->events))
+	if (!kfifo_is_empty_spinlocked_noirqsave(&le->events, &le->wait.lock))
 		events = EPOLLIN | EPOLLRDNORM;
 
 	return events;
@@ -830,43 +827,52 @@ static ssize_t lineevent_read(struct file *filep,
 			      loff_t *f_ps)
 {
 	struct lineevent_state *le = filep->private_data;
-	unsigned int copied;
+	struct gpioevent_data event;
+	ssize_t bytes_read = 0;
 	int ret;
 
-	if (count < sizeof(struct gpioevent_data))
+	if (count < sizeof(event))
 		return -EINVAL;
 
 	do {
+		spin_lock(&le->wait.lock);
 		if (kfifo_is_empty(&le->events)) {
-			if (filep->f_flags & O_NONBLOCK)
+			if (bytes_read) {
+				spin_unlock(&le->wait.lock);
+				return bytes_read;
+			}
+
+			if (filep->f_flags & O_NONBLOCK) {
+				spin_unlock(&le->wait.lock);
 				return -EAGAIN;
+			}
 
-			ret = wait_event_interruptible(le->wait,
+			ret = wait_event_interruptible_locked(le->wait,
 					!kfifo_is_empty(&le->events));
-			if (ret)
+			if (ret) {
+				spin_unlock(&le->wait.lock);
 				return ret;
+			}
 		}
 
-		if (mutex_lock_interruptible(&le->read_lock))
-			return -ERESTARTSYS;
-		ret = kfifo_to_user(&le->events, buf, count, &copied);
-		mutex_unlock(&le->read_lock);
-
-		if (ret)
-			return ret;
-
-		/*
-		 * If we couldn't read anything from the fifo (a different
-		 * thread might have been faster) we either return -EAGAIN if
-		 * the file descriptor is non-blocking, otherwise we go back to
-		 * sleep and wait for more data to arrive.
-		 */
-		if (copied == 0 && (filep->f_flags & O_NONBLOCK))
-			return -EAGAIN;
+		ret = kfifo_out(&le->events, &event, 1);
+		spin_unlock(&le->wait.lock);
+		if (ret != 1) {
+			/*
+			 * This should never happen - we were holding the lock
+			 * from the moment we learned the fifo is no longer
+			 * empty until now.
+			 */
+			ret = -EIO;
+			break;
+		}
 
-	} while (copied == 0);
+		if (copy_to_user(buf + bytes_read, &event, sizeof(event)))
+			return -EFAULT;
+		bytes_read += sizeof(event);
+	} while (count >= bytes_read + sizeof(event));
 
-	return copied;
+	return bytes_read;
 }
 
 static int lineevent_release(struct inode *inode, struct file *filep)
@@ -968,7 +974,8 @@ static irqreturn_t lineevent_irq_thread(int irq, void *p)
 		return IRQ_NONE;
 	}
 
-	ret = kfifo_put(&le->events, ge);
+	ret = kfifo_in_spinlocked_noirqsave(&le->events, &ge,
+					    1, &le->wait.lock);
 	if (ret)
 		wake_up_poll(&le->wait, EPOLLIN);
 
@@ -1083,7 +1090,6 @@ static int lineevent_create(struct gpio_device *gdev, void __user *ip)
 
 	INIT_KFIFO(le->events);
 	init_waitqueue_head(&le->wait);
-	mutex_init(&le->read_lock);
 
 	/* Request a thread to read the events */
 	ret = request_threaded_irq(le->irq,
-- 
2.25.0


  parent reply	other threads:[~2020-02-11  9:19 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-11  9:19 [RESEND PATCH v6 0/7] gpiolib: add an ioctl() for monitoring line status changes Bartosz Golaszewski
2020-02-11  9:19 ` [RESEND PATCH v6 1/7] kfifo: provide noirqsave variants of spinlocked in and out helpers Bartosz Golaszewski
2020-02-11  9:19 ` [RESEND PATCH v6 2/7] kfifo: provide kfifo_is_empty_spinlocked() Bartosz Golaszewski
2020-02-11  9:19 ` Bartosz Golaszewski [this message]
2020-02-11  9:19 ` [RESEND PATCH v6 4/7] gpiolib: emit a debug message when adding events to a full kfifo Bartosz Golaszewski
2020-02-11  9:19 ` [RESEND PATCH v6 5/7] gpiolib: provide a dedicated function for setting lineinfo Bartosz Golaszewski
2020-03-16 16:20   ` Geert Uytterhoeven
2020-03-17 13:40     ` Bartosz Golaszewski
2020-02-11  9:19 ` [RESEND PATCH v6 6/7] gpiolib: add new ioctl() for monitoring changes in line info Bartosz Golaszewski
2020-02-12 10:47   ` Linus Walleij
2020-02-12 11:00     ` Bartosz Golaszewski
2020-02-20 15:03       ` Linus Walleij
2020-02-20 15:06         ` Bartosz Golaszewski
2020-02-11  9:19 ` [RESEND PATCH v6 7/7] tools: gpio: implement gpio-watch Bartosz Golaszewski
2020-02-12 10:49 ` [RESEND PATCH v6 0/7] gpiolib: add an ioctl() for monitoring line status changes Linus Walleij

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=20200211091937.29558-4-brgl@bgdev.pl \
    --to=brgl@bgdev.pl \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=bgolaszewski@baylibre.com \
    --cc=linus.walleij@linaro.org \
    --cc=linux-gpio@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=warthog618@gmail.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;
as well as URLs for NNTP newsgroup(s).